The Digital Transformation of Consumer Experiences – Trends We See Accelerating

The secular shift from physical to digital commerce was well underway before the global pandemic. What we have seen in the past six months is an acceleration of this trend which is further blurring an already blurred line between the two.

When retail stores were forced to close, brands like Lululemon turned their physical stores into fulfillment nodes for their online business, with store employees coming in to pick, pack, and ship from store inventory. At retailers like Target and WalMart, curbside pickup and store-based delivery quickly became mainstream ways for customers to shop. At grocers like Kroger and QFC, Instacart and grocers’ own shelf-picking and delivery services overlaid digital transactions on physical commerce infrastructure. There is no question that the shift we have seen in 2020 will create opportunities for years to come, both in accelerating more transactions from physical to digital and in leveraging the existing physical retail infrastructure to support digital use cases. These are markets where Madrona has been investing for 25 years, and they continue to be focus areas for us. The commerce-related investment themes we talked about in 2019 are increasingly relevant today. This post updates our thinking on how retail is undergoing a significant shift – and outlines the areas we think are ripe for innovation, and for investment.

Digital+Physical

As we think about the intersection of the digital and physical more broadly, the shift of in-person experiences to remote/digital has been accelerated by the global pandemic. From ceremonies and celebrations, to social outings to health visits and entertainment, the activities we once felt (very strongly!) could only happen in-person are shifting online. Initially this was out of necessity, but over time we believe online will be a significant portion of the market, as the quality and convenience are demonstrated and becomes socially acceptable. In this post, we’ll cover four markets where this dynamic applies: Telehealth, Shopping (product discovery), Fitness, and Entertainment.

Fitness and healthcare are markets where we’ve been tracking the shift from in-person (physical) to digital and have been hunting for opportunities. Live (in-person) entertainment and product discovery (a subset of the offline shopping experience), are newer verticals where we are beginning to see or expect to see major shifts of attention and dollars to digital. We believe there are compelling opportunities for new companies to be built and existing companies to accelerate growth and share across all four markets.

Telehealth

As we have been unwilling or unable to visit our doctors’ and medical professionals’ offices during Covid, large segments of the medical industry have been forced to shift to a digital-first mode of care. This means seeing and understanding patients via high fidelity video, connected sensors, and other technologies to triage and diagnose potential issues as well as perform on-going monitoring and support.

The magnitude of this shift and the opportunity it creates cannot be overstated. The healthcare market is enormous: US total market size in 2019 was approximately $2.7 trillion. If you strip out hospital care, pharmaceuticals, and nursing/home health, you’re left with about a $900 billion health services market. We believe a large portion of this $900 billion can ultimately be delivered remotely through digital means.

One vertical where we believe there is opportunity for multi-billion dollar businesses is mental health. Approximately 46 million Americans report having mental health issues, with only 43% of those reporting having received treatment. A recent survey of therapists found that during Covid, 75% of therapists had shifted to digital-only therapy sessions while 16% were doing a combination of digital and in-person.

Mental health is a compelling category for a number of reasons. First, both 1:1 and group therapy sessions can translate well to video. Second, by relaxing constraints like geographic proximity between patient and therapist, scheduling is easier, more patients can be served and no “office” means lower overhead and lower cost of care. Additionally, for the youth segment of the mental health industry where conditions such as anxiety and depression are skyrocketing, digital is a seamless and native means of communication. Finally, industry studies, while early, suggest parity in effectiveness of in-person and virtual mental health therapy delivery.

The structural challenges around areas like insurance reimbursement/coverage and HIPAA compliance will be resolved over time — we have seen many fast-tracked solutions and progress in recent months — and it’s likely that the ultimate model of delivery for mental health will be a hybrid of in-person and digital. But the prospect of a digital-first delivery method for health services in general and mental health services, specifically, is a compelling opportunity for health consumers and the industry overall. It’s still very early days for companies being built in this category, but there’s no shortage of opportunities for innovation around technology, customer experience, and business models in the delivery of digital/remote mental health services into the mainstream.

Product Discovery (Shopping)

Much of the magic of physical in-store shopping is how it enables discovery, learning, and commerce in a single place, with the core ingredient being people. We find expertise every time we go to the store. For example, want to start a vegetable garden and don’t know where to begin? Head down to your favorite specialty garden shop or big box home improvement retailer to get advice from an expert. Need a new pair of running shoes, but you aren’t sure which ones work best with your running style? Head down to the specialty running store and they’ll ask some questions and watch you run. Want to take the kids camping but have no idea what you need? Head over to REI and an employee will outfit you head to toe based on the trek you will be taking. Physical retail is magical because it combines the expert knowledge of the human working at the store with the transaction itself. Conversely on the web, while there is no lack of content and inventory, the process of discovery, learning, and aided purchasing are broken. For example, Amazon has a massive amount of content but it’s not really a good place to discover new products or get expertise in a given category. Even the websites of the specialty retailers with incredible product and category knowledge really don’t get the job done – because those businesses are still oriented toward the in-person expertise model.

In a world in which we are taking fewer trips to the retail store, and the core physical retail business model is in peril, it begs the question: How do we create the digital analog of a more complete shopping experience? How do we bring the wealth of retail employee expertise and experience online? At Madrona, we see this as a big opportunity and believe that successful solutions will offer a magical combination of discovery, learning, and commerce (with a bit of entertainment), built into a highly integrated and seamless customer experience.

What does this look like? We’re not sure we’ve seen it yet, but we have a few ideas as to where there is opportunity for innovation. In the legacy experiences category, TV shopping through networks such as QVC and HSN is a great example. Discovery is core to that experience, with a curated set of products found all over the globe by the QVC/HSN sourcing teams as well as a host who walks viewers through the features and benefits of products with finely honed sales pitches, informed by data, to get you to purchase. “Buy now!” has shifted significantly from a phone number to navigate to the URL, but there is a direct call to action and connection between the content and the commerce. QVC/HSN did $11 billion in TV-driven retail revenue last year … while not the size of Amazon, it is a significant business. We think someone ought to re-invent HSN/QVC for the digitally native buyer.

Two new experiences we find fascinating were built outside of the U.S. market, but portend opportunities anywhere. One is ShopShops, a Chinese e-commerce company that uses live video to help domestic Chinese consumers shop retail shelves for the hottest brands in L.A., New York, and Miami (and soon, many other places). Like QVC, ShopShops hosts showcase merchandise and provide a curated retail experience, but they do so in partnership with physical retailers and brands, bringing products and knowledge together, and providing an integrated way to buy. In a world in which consumers want to see what’s new, but don’t want to venture out to the retail stores themselves (or the retail stores are halfway around the world), this is a novel concept that uses existing retail infrastructure and live video to enable discovery, learning, entertainment, and commerce, all over the world. Another is BulBul, an India-based company that enables anyone to build a live streaming channel through which they can promote and sell products. This is sort of like a crowdsourced QVC; instead of a managed production, BulBul is a marketplace model. We know that marketplace models don’t always deliver pristine customer experiences but leveraging the expertise of the crowd for any product category is an interesting idea.

We believe that elements of all of these can drive the translation of the discovery, learning, and commerce that happens in-store to a digital experience.

Fitness

While previously large group fitness classes were a popular way to get a workout and be social, these classes have been replaced with home gyms and outdoor workouts since Covid started. Even 1:1 sessions with trainers, which typically take place in gyms with shared equipment, have faced similar declines due to the nature of the in-person fitness environment. The market for individual and group training is enormous. The quality of the instructor and their ability to motivate you, watch your form, and give feedback is easy to do in-person. Showing up for the other members of your class and encouraging each other to achieve your goals is also effective in-person. So, how do these translate digitally?

Peloton demonstrated (prior to Covid, but also accelerated by Covid) the effectiveness of designing a digital experience around and into a physical piece of equipment – the spin bike. We’ve seen from Peloton and others the extension of this model of incorporating live instruction and community into other pieces of equipment such as the treadmill, the rowing machine, the multi-trainer, and even the gym mirror. Some of these hardware-centered models will work, although being first to market is key because it’s unclear how many Peloton-like offerings can exist (both literally and financially) in a given household. It may be that, like in the wearables market, there’s one new category creator (Fitbit) and some large incumbents (like Apple) who build or buy their way in, leveraging their power of product development, marketing, and distribution.

But… is equipment-centered the only opportunity to deliver a digital one-to-one or one-to-many fitness experience that both lives up to its analog/in-person counterpart, and provides a compelling business model? We believe no, and that there is a lot of room for this market to grow.

Entertainment

Professional sports are coming back (kind of) … but without fans in the stands. The business model for live professional sports has a huge in-person component: ticket sales, parking, in-venue merchandise sales and concessions comprise a huge percentage of professional sports revenue. But the other big component for professional sports is broadcast revenue, which exists with fans or without. While broadcast revenues won’t make up for the shortfall from in-venue experiences, professional sports will find a way to survive. The movie industry should hold up alright as well (although perhaps not the theaters), as new movies go straight to over-the-top and subscription video services.

But what happens to other in-person entertainment experiences without an existing broadcast or video component and revenue stream? Theater, comedy, and music all depend heavily on ticket sales and ancillary revenues generated from in-venue and sponsorship experiences.

We are particularly interested in professional music given the sheer magnitude of the market. Live Nation / TicketMaster is facing an existential threat with revenues down a whopping 98% at the end of June 2020 as compared to the end of June 2019. The market needs radical rethinking on how to diversify and move part of the consumer experience to digital. Over the past few months, it has become clear that Live Nation and Ticketmaster’s strategy is to wait-it-out instead of putting serious resources into digital. Why isn’t the music industry doing more to drive innovation in content delivery and business model innovation for live experiences?

An argument we’ve heard from several in the industry is that certain elements of the live concert experience simply don’t translate to digital. The group of friends you go to the concert with, the pre-party in the parking lot or at the bar or restaurant, buying and wearing the shirt in-venue, singing back the song as part of the crowd, the energy of thousands or tens of thousands of bodies swaying to the music, the production of the show itself. The sounds, the smells, the experiences in-venue simply can’t translate to digital, and we shouldn’t try.

While we think there’s some truth to that argument, we also think there is opportunity for new digital experiences to be good enough to at least capture some of the massive market at stake. The show must go on for artists to survive and sponsors and brands are searching for ways to reach an audience during the current live events vacuum.

Is there an opportunity to translate live music experiences to digital with a monetization model that works? We believe there could be, and that it’s going to take a lot of experimentation and innovation to figure it out. “Twitch for live music” probably isn’t going to work (just to use an analog; this is not a critique of Twitch). There’s some interesting experimentation with virtual concerts, including leveraging artists’ distribution (e.g., Travis Scott on Fortnite). There’s live video on Instagram Live and sites like LiveConcertsStream, and, of course, there’s always Zoom’s video conference platform which can be repurposed for events like these. But do we sell tickets, or do we accept tips? Do artists perform from their homes, or use small studios and production venues with high fidelity video and sound? Sports like boxing have found a way to meld an in-person experience with a remote (TV) experience, where pay-per-view revenue is a larger source of revenue than in-person revenue. Can pay-per-view (or something like that) be additive to the live music model?

In Conclusion

Now is the time for experimentation and testing, not just for music, but for all of these categories. We are excited to see new business models that bridge the digital and physical world and enable new experiences along the way. From fitness to music to shopping to mental health, the opportunities to build are endless and we are eager to meet entrepreneurs innovating in these categories.

Introducing a New Software Stack – The Essential Revenue Stack

From working with our 90+ portfolio companies and their customers, as well as from frequent conversations with enterprise leaders, we have observed a set of software services emerge and evolve to become best practice for revenue teams. This set of services – call it the “revenue stack” – is used by sales, marketing, and growth teams to identify and manage their prospects and revenue.

The evolution of this revenue stack started long before anyone had ever heard the word coronavirus, but now the stakes are even higher as the pandemic has accelerated this evolution into a race. Revenue teams across the country have been forced to change their tactics and tools in the blink of an eye in order to adapt to this new normal – one in which they needed to learn how to sell in not only an all-digital world but also an all remote one where teams are dispersed more than ever before. The modern “remote-virtual-digital”-enabled revenue team has a new urgency for modern technology that equips them to be just as – and perhaps even more – productive than their pre-coronavirus baseline. We have seen a core combination of solutions emerge as best-in-class to help these virtual teams be most successful. Winners are being made by the Directors of Revenue Operations, VPs of Revenue Operations, and Chief Revenue Officers (CROs) who are fast adopters of what we like to call The Essential Revenue Software Stack.

In this Essential Revenue Software Stack, we see four necessary core capabilities – all critically interconnected. The four core capabilities are:

  1. Revenue Enablement
  2. Sales Engagement
  3. Conversational Intelligence
  4. Revenue Operations

These capabilities run on top of three foundational technologies that most growth-oriented companies already use – Agreement Management, CRM, and Communications. We will dive into these core capabilities, the emerging leaders in each, and provide general guidance on how to get started.

Revenue Enablement

Revenue Enablement – what we see as the core of the Essential Revenue Stack because of how the other components rely on it to tie everything together – is a relatively new concept that has quickly become a standard best practice over the last few years.

It is where Sales, Marketing, and Revenue leaders want to enable every customer-facing employee to be smarter in how they work with customers. With the usage of Revenue Enablement tools, reps can more nimbly utilize existing marketing and product content to support their conversations as well as augment asynchronous training with real-time on-the-job learning in order to deliver a consistently exceptional customer experience.
We see Revenue Enablement as the centerpiece of the modern revenue stack because it is the hub with which other systems interface. For instance, Sales Engagement systems like Outreach pull core prospecting content and communication templates from a Revenue Enablement system and share back activity and analytics so that there is a comprehensive picture of the effectiveness of engagement activities.

Great Revenue Enablement solutions seamlessly deliver the full spectrum of “enablement”: AI-driven search and predictive content recommendations, guidance that helps reps understand exactly what to “know, say, show, and do,” and real-time training to reinforce core concepts at the point of engagement. For years, companies’ go-to-market (GTM) approach and key “sales plays” have existed only in PowerPoint or a strategy document. Now, companies can schematize their sales motion, breaking it into discrete steps that enable them to measure, manage, and optimize it.

We at Madrona Venture Group have witnessed the evolution of the Revenue Stack and the development of the Revenue Enablement function as Highspot investors from the earliest days. Highspot has emerged as a category leader due to its simple and elegant user experience, powerful search and breadth of features. Other companies such as Showpad and Seismic also have strong offerings and customer rosters.

Sales Engagement

Some think of Sales Engagement as an intelligent e-mail cannon and analysis engine on steroids. While in reality, it is much more. Consider these examples: How can I communicate with prospects in a way that is both personalized and efficient? How do I make my outbound sales reps more productive and enable them to respond more quickly to leads? What tools can help me with account-based marketing? What happened to that email you sent out to one of your sales prospects?

Now, take these questions and multiply them by a hundred, or even a thousand: How do you personalize a multi-touch nurture campaign at scale while managing and automating outreach to many different business personas across various industry segments? Uh-oh. Suddenly, it gets very complicated.
What Sales Engagement comes down to is the critical understanding of sending the right information to the right customer, and then (and only then) being able to track which elements of that information worked (e.g., led to clicks, conversations, and conversions)… and, finally, helping your reps do more of that.

We see Outreach as the clear leader here, based in Seattle, with SalesLoft as the #2. Outreach in particular is investing considerably in adding additional intelligence and ML to their offering to increase automation and improve outcomes.

Conversational Intelligence

Conversational Intelligence tools help revenue teams – and especially revenue leadership – monitor and interpret the conversations that are happening between their reps and prospects. This analysis creates a stronger learning loop for revenue teams. Just as athletes and coaches watch game film before and after games and during practice, reps and leadership can use these tools to listen and learn from sales calls, make improvements to tactics and strategies, and determine how to best coach reps over time.

For example, it has become commonplace for some of the CEOs in our portfolio to listen to sales call recordings on their commutes or during workouts, etc. It is an effective and efficient way for them to keep a finger on the pulse when it comes to understanding the voice of the customer, guide the fine-tuning of the company’s and product’s messaging and positioning, and learn what good reps sound like. Importantly, from a CRO perspective, it also helps reps hone in on the most effective messaging and sales tactics that ultimately lead to higher “close” rates and faster topline growth.

Gong has emerged as the market leader, but a number of our portfolio companies use Chorus as well.

Revenue Operations

RevOps is all about creating a better-informed view of the sales funnel through helping drive alignment across sales, marketing, and customer service teams, data, and processes. According to Clari, “RevOps is the strategic convergence of sales, marketing, and customer success to drive full-funnel accountability across the revenue engine.” Most importantly – it creates a stronger ability for CROs and Sales Leaders to predict future revenues with greater confidence through the ability to more easily understand their data and create powerful views of revenue for the business.
There are many new players in the market that are adding these kinds of RevOps/operational intelligence layers on top of data in Salesforce. They provide a deeper and easier-to-access understanding of core business metrics and, most importantly, more accurate revenue forecasting. Through our work with a broad range of enterprise customers, we have seen our portfolio company, Clari, emerge as a leader in this area. We have seen CEOs and boards come to rely on “the Clari number” as a reliable and indispensable indicator for quarterly forecasts.

Another important component of Revenue Operations (sometimes referred to as Revenue Intelligence) is intelligently capturing and synchronizing data across your revenue systems and ensuring your CRM contains ground truth, particularly as it relates to contacts (leads and customers). In other words, ensuring your systems of engagement are harmonized with your system of record. In addition to Clari, we are seeing Salesforce Einstein and People.ai as players in this area.

Where Should I Start?

While there isn’t a standard approach to where you should begin modernizing your revenue stack, there are some basic steps you can take to know where to focus your approach. First, make sure you have the right foundational technologies.

Agreement Management

Increasingly, contracts and agreements are no longer “gathering dust” (either physically or digitally) – they are instead being used to provide scalable, intelligent triggers for business events such as contract renewals and upselling.

The culmination of a sale is getting a signature on the dotted line. But when a company has hundreds of customers and thousands of contracts, keeping track of specific account details and when each comes up for renewal has historically been a manual and time-consuming process.

There are already a number of solutions creating basic contract execution and signature capabilities, DocuSign being the industry standard. Recently DocuSign has been pushing to expand into end-to-end agreement management and what they have coined the “agreement cloud.” Other players such as Apptus (which recently acquired Conga) offer related capabilities such as quote-to-cash, contract management, and document generation. Then there are pure-play contract management offerings such as iCertis (a Seattle-based leader in enterprise contract lifecycle management), Lexion (another Madrona portfolio company) which is a machine learning-centered offering that uses NLP to automatically extract key contract terms, and IronClad which can incorporate sales-related workflows.

CRM & Communication

Foundational to not only the revenue-generating side of the business but also the rest of the organization, are two core technology capabilities: CRM and modern communication tools. Let’s quickly break these down:

  • CRM: Salesforce is the de facto standard used by many organizations we see. It acts as the system of record for all customer relationship management (CRM) information, both feeding and ingesting from the other software systems in the revenue stack. It is also rolling out a new Salesforce Anywhere App in the second half of 2020, with new tools designed specifically for the remote-virtual-digital workforce.
  • Communication: These include both asynchronous text-based communication and real-time video collaboration. The best in class text-based communication tool is Slack, with organizations using Slack to chat both internally as well as externally with prospects and customers. As is widely recognized, Zoom has become the standard for video communication, providing the best performance, most complete feature set, and broadest set of integrations. We have started to see more companies using Microsoft Teams for both text-based chat and video. While not as good at threaded chat as Slack or as good at video as Zoom, the fact that it provides both, is slickly integrated with Outlook, and is free with Office365, has led the many companies that already use Microsoft for email to adopt Teams.

Next, take inventory of the core capabilities you’ve already adopted across the revenue stack, using this article as your guide. Many organizations have already adopted some pieces of the revenue stack but haven’t yet integrated these technologies to their full potential. Others still need to take the first steps to modernize. Either way, we recommend as a first-order priority to start with evaluating the current state-of-play of your Revenue Enablement function, and then move on to Sales Engagement, Conversational Intelligence, and Revenue Operations. There are literally hundreds of different software tools targeting sales teams, so be careful not to be distracted by “tool overload” constantly looking at the next shiny new object. We find companies that focus on these four core capabilities maximize their success.

In Summary

We believe we are living in a new normal. “Remote-virtual-digital” work is going to be a key part of the playbook forever, and The Essential Revenue Software Stack is the best way to set up your teams for success. The power of this stack – Revenue Enablement, Sales Engagement, Conversational Intelligence, and Revenue Operations, integrated on top of your core CRM, communication tools, and Agreement Management – can drive significant, measurable ROI for your business. If you don’t use it, you’re at a disadvantage.

*****

This was first published by TechCrunch.

Our Investment in Coda & The Future of Work

Today we are excited to announce our investment in Coda.

Recently, we wrote some blog posts about our core investment themes. Specifically, two of the themes relating to Coda are the Future of Work, Workforce and Workplace as well as Low code/No Code development platforms.

There are several market trends that drive our thesis on these areas.

  1. Companies want to move faster and be more agile
  2. Cloud-native apps make it easier to access data
  3. More “makers” than ever before
  4. Multi-player collaboration and digital-first workflows

These trends together are driving innovation in how people are reimagining what tools people need to be effective and successful in today’s day and age.

The team at Coda recognized these trends when they first began working on Coda and have been building a fantastic product to serve these market needs.

The Coda team saw these trends and through their product have fundamentally reimagined how we work. They recognized that we move from application to application, and when applications talk to each other to share data and be dynamic, that is where the work gets done. So, they started with that as a core design point.

Coda is a single canvas that brings together the best of documents, spreadsheets, databases and applications. The flexibility of Coda enables users to become ‘makers’, authoring dynamic documents where the lines between a document and application are blurring and becoming one and the same.

Coda envisions a future where documents are alive and interactive – enabling users to interact with data, to interact with systems, and to automate previously manual processes. Coda imagines a world where everybody can be a “maker” and can use Coda to express what they want to and collaborate with others seamlessly.

We believe this is a massive opportunity. The “Maker Generation” has rapidly adopted new platforms and tools to solve problems, build products, and start new businesses, and we have seen many examples of this in other industries – from producing channels on YouTube, to building web sites in WordPress, or developing new games on Roblox.

Coda’s team believes they can unlock the power of software development for the Maker Generation by providing them with the platform and building blocks to build apps that look like docs, and we are already seeing thousands of users building these apps on Coda today.

We are particularly committed believers because we are Coda users. As investors we make decisions on whether to invest in an idea, founder or company together and on many different factors. We built one of the tools we use to start conversations and look at challenges, in Coda. Earlier this year we published this as a document in the Coda Doc Gallery for anyone to use.

Beyond the large market opportunity, we are also thrilled to back this world-class team. We have known Shishir Mehrotra for years since Microsoft where he worked on SQL Server, Windows, and Office before eventually moving down to the Bay Area to join Google as YouTube’s VP of Product, Engineering, and UX. Shishir’s cofounder and CTO, Alex DeNeui is also a world-class technology leader in real-time collaboration tools, and his previous company, DocVerse, was acquired by Google.

With the combination of a fantastic team, great product, and a massive market opportunity, we are looking forward to joining Coda in this journey to reimagine the future of productivity and collaboration. Coda’s blog on their funding is here.

AND we talked to Shishir about his journey to start Coda for our newest Founded and Funded Podcast – you can listen here!

 

Podcast Transcript

 

Erika: [00:00:11] Welcome to Founded and Funded. I’m Erika Shaffer with Madrona Venture Group. And today we’re really excited to bring you Shishir Mehrotra. The founder and CEO of Coda. Coda announced a new funding round today of $80 million, which we participated in and in this conversation with Madrona managing director Soma and Madrona, senior associate Elisa La Cava, Shishir talks about what Coda is, how it got founded, what you really need to start a company in his view and the very interesting journey that Coda has taken to release their first product, which was five years in the making. Before founding Coda show this year started his company. Worked at Microsoft for many years where he knew Soma. And then went to run YouTube for Google. Let’s pick up here, where Soma introduces the conversation.

Soma: [00:01:12] Hello, everybody. I’m very excited to be here today to talk about one of our core investment themes, the future of work with the founder and CEO of Coda Shishir Mehrotra. Welcome to Shishir.

Shishir: [00:01:27] Hi, Soma. It’s nice to be here. I was just to comment that you’re one of the few interviewers who can pronounce my name so clearly, okay.

Soma: [00:01:35] Yep. It’s another advantage of coming from the same part of the world. Yeah, but Shishir, really excited to have you on this podcast with us today and as we are very thrilled to be on the Coda journey with you as investors in the funding round that you announced earlier today.

Shishir : [00:01:50] Yeah. Welcome aboard.

Soma: [00:01:51] Thank you. We absolutely believe the Coda is one of the companies leading the charge on enabling what we call the future of work. The thing that is most impressive to me about Coda, when I think about it is how you all are reimagining how the world of documents and applications can come together and goes way beyond what people have access to in terms of productivity tools and collaboration tools that they are used to today.

So that’s been very exciting for us. And the future of work is an area that we’ve been investing for many years now, but we believe that the pace of innovation and more importantly, the demand from customers for adoption of great communication, collaboration tools has dramatically accelerated since the onset of Covid.

We are also going to have with us Elisa La Cava, one of my colleagues at Madrona and she’s been part of the team that has helped us formulate a lot of our thinking on the future of work. So I thought, Shishir, that we’d get started off with your personal journey, talking with you, talking a little bit about your personal journey, leading up to both founding and stuff in Coda, you came out of MIT with your undergrad and right off the bat, you founded a company called Centrata, which was built around that. How do you describe the entrepreneurial bug that bit you at that early stage in your sort of career as you are just coming out of school? And what do you think it takes for a person today to cross over into entrepreneurship, challenge themselves and decide to create something that is going to be phenomenally better for the world at large?

Shishir Mehrotra: [00:03:26] I think it’s a great question. Maybe as a piece of background, this is my second time founding a company , I’ve been involved with startups for a long time, but, directly second time founding a company.

And, I think people are addicted to starting companies. I’m sure. I’m sure you invest in some people where every idea that I think of is formulated as a startup. I’m not one of those people I’m quite comfortable with large environment, small environments, like you I’ve had a chance to work in both of those contexts.

And so it gives me a little bit of appreciation for entrepreneurs and what makes that unique. And I think. When I was at Google before, a very common interaction with somebody would come to me and say, I want to leave to go start a company. And each of us, I’m sure you had this in your past roles as well.

We just developed our viewpoint on how to have these conversations. And I settled on two questions and generally these questions were asked a little bit as almost as a deterrent in some ways, the two questions were. If somebody says they want to start a company, I would ask number one.

Do you have an idea you can’t imagine not working on and number two, do you have a person you can’t imagine not working with? And inevitably people would answer yes. To one of the questions and no to the other. And then we’d have a conversation about, why magic hits when these two things come together.

And they talk about we’ll have this great idea, but I haven’t been able to convince anybody else to do it, or they would talk about how they have a perfect partner, but they haven’t settled on an idea yet. And, I found that the, for entrepreneurship to hit that sort of magic has to hit at that right moment.

So those are the two questions that I’ve developed over time as a kind of litmus test of, should you start a company? And actually in my own journey with Coda, I wasn’t actually trying to start a company. I was fairly sure I would have continued on at Google, running large teams. And for me, these two questions, all of a sudden they answered yes. And I just, couldn’t not start a company. And I often describe entrepreneurship, not as a gift, but as a curse. And at that point, you just can’t think about anything else and everything else seems small. Everything else seems not worth doing. That’s when you, I think start a company and jump all the way in.

Soma: [00:05:30] It’s pretty impressive to see the journey you’ve gone through Shishir, but let me hand it over to Elisa to ask you the next set of questions.

Elisa La Cava: [00:05:38] So after the Centrata experience, I know you did a tour of duty, about six years at Microsoft and another six at Google before founding Coda.

And what I’m curious is what was different for you? If anything, this time around when you started Coda versus when you started Centrata during the height of the .com era, what was different for you in terms of the conditions to start a company, but also your entrepreneurial mindset and drive and determination to build something new.

Shishir Mehrotra: [00:06:13] very interesting question. At 15 years apart in that cycle, my situation the starting of the companies could not be more different. Cenrata I was coming out of school, we were converting my graduate work into a company. One fun story was the way the financing happened.

The company was in Toronto and was funded by, a guy named Vinod Khosla. We had been trying to raise money for nine months, flying back and forth from Boston to California. And no real success in doing it. And I get this email from Vinod and, it says, I read your business plan.

I’d like to, I’d like to fund your business. Vinod, is pretty direct, so his emails are direct and short, they’re full of misspellings. It’s like half the words or half the words are misspelled. And so I get this email and, this is, it’s 2000, this is a timeframe when there was no Wikipedia, there was no LinkedIn.

So I, this is a little embarrassing to say, but I had no idea who Vinod was. And so I write one of my, angel. So I had this angel, who had promised to put a half a million bucks into the company, but hadn’t actually done it yet. And so I write him and I say, do you think this email is real?

Is this like spam? Is this what, it’s full of misspellings. I’m not sure who it is. And so the angel says to me,

Elisa La Cava: [00:07:18] Right and you and you hadn’t, you hadn’t spoken yet. He had just looked at your business plan without even talking to you.

Shishir Mehrotra: [00:07:23] And it’s like, what are the chances? That somebody wants to fund the business.

So the angel says to me why don’t you walk down the street to the bookstore, the college bookstore and go to the magazine rack, see the person whose face is on all the magazines. That’s Vinod Khosla because that’s when he was at the top of his game. The funny part starts when I come back to my dorm room and my CFO in quotes was my, housemate.

I’m walking back into the house and I’m excited about this and he stops me and I said, I have something important to share. And he says, Oh yeah, I was actually gonna ask $500,000 just showed up in the bank account. Like what, where did that come from? This angel had heard this and wired the money in.

The first experience of starting Centrata was, we didn’t know what we were doing. I didn’t know who the financers were. You’re just figuring everything out from, from scratch. And I think there’s a level of blind determination. You’re not bogged down by the reality of the world. Coda, I got started, we raised our first round of financing in a weekend. I had a much better sense of what we were doing. I knew what terms to ask for. and so it’s like dramatically different that way.

But what I’d say about that was similar was at these two incredibly different moments of my life. One where, I didn’t really know what we were doing. And the other one where, I had a better sense of how this whole process works. The similarity, I think was that same level of ridiculous conviction on an idea that honestly, everybody else around us thought was weird and I wasn’t quite sure what to make of it. And I think that similarity and that determination, you can spot it, the eyes of every entrepreneur at that moment. They just don’t understand why you don’t get it.

It’s just so obvious. Like it says, obviously going to work and that’s what you have to have because you’re going to get every type of you’re going to get every type of no, why don’t you go get a real job, getting all different versions of that.

And when it works, it’s a particularly exciting, but that’s some similarities and some dramatic differences.

Elisa La Cava: [00:09:12] I love that though. You’re talking about that determination and it sounds like with Coda, you kept thinking about it and thinking about it. And then at one point you just thought, Oh, like I have to do this.

There’s no alternative.

Shishir Mehrotra: [00:09:25] Coda is one of those products where first off, when you can picture it just feels that’s obviously how it should be.

And that, that seems like really clear. Why would anybody bother building it any other way? The other thing that happens with Coda is it’s like this meta product where almost every idea that was pitched to me, I could picture that product, every idea I heard was, Oh yeah, I should just build that on this platform.

Like that. that’s exactly what we should go do. And so it became this sort of, every idea felt small compared to this thing. Now, the flip side, I call my parents, my wife’s on, they all look at me and say, what’s wrong with Microsoft Office? Like, why do you want to go do that? That seems silly.

So you had to, square those two pieces away. And of course, for the, when you cross that entrepreneurship hill and you’re in that convicted, you can’t imagine not working on this thing. You can’t imagine it not working now. Everybody else’s lack of faith actually, emboldens you and you get even more determined.

And even when you get even more certain that, if I don’t do this, then nobody will do it and then I’ll feel really bad, and then, I think Jeff Bezos was talking a lot about regret minimization framework. Yeah. Then you’re going to really feel like, man, if I had just done that, then the world would have this new product , and it would have worked out in this different way.

And so I think that the, that determination is very similar.

Soma: [00:10:42] That’s a fantastic story, Shishir. You already started talking a little bit about Coda, but I thought, let me pause for a second and hear from you. How would you describe, what Coda is all about?

Shishir Mehrotra: [00:10:53] Yeah. I think our users would describe Coda as an all-in-one interactive document and what they would probably say is it blends the best parts of documents, spreadsheets, presentations, and applications together into one new surface.

Our promise is that it allows anyone to make a doc as powerful as an app. And that’s a, that’s the bold promise and something that we’re fairly committed to the, I think that if you step back for a moment, Coda was formed with two primary observations of the world. The first observation is that the world runs on docs, not apps. And that if you were to look yeah, that any team, business, family, individual, and say, what do you use to run yourself, your team, your business, so on. They’ll probably rattle off a set of applications they use, and a lot of packaged applications and I have the CRM system and this inventory system and this task system, and, so on.

But then if you watch them all day long and just stand behind their desk and see what they’re working on, you’ll see them in documents, spreadsheets, presentations, and communication tools all day long. And this observation was pretty stark when I worked at Microsoft on the office team, so that was pretty stark then, but it was particularly vivid for me at Google.

I got to Google 2008, right when Google docs came out and it was transforming the way we ran our businesses. And so YouTube, 2008, we basically ran everything on Google Docs, Google Sheets, and Google Sites. And, there was some kind of extreme examples of this, I have to pick one crazy example. If you hit flag on a YouTube video back in that 2008, 2009 period, it would create a row in a spreadsheet on an ops person’s desk. And that was how, that’s how pervasive this was. And for a lot of people that sounded crazy. But for me, it was part of this observation, that docs, not apps run the world and I think people saw that as a weakness, I saw that as a strength, it gave us complete agility. It meant that when we wanted to change how we did planning, we could do it instantly.

And as the world of how we thought about flagging and content moderation, so on evolved we had total control over it. So this kind of observation, number one, docs not apps run the world. The second observation is that those surfaces document, spreadsheets, presentations haven’t fundamentally changed in over 40 years.

And there’s a running joke in the company that if Austin Powers popped out of his freezing chamber, he wouldn’t know what clothes to wear. He wouldn’t know what music to listen to, but he would know how to work a document, a spreadsheet and presentation because none of them have fundamentally changed since the 1970s.

And this is, if you go back to WordStar Harvard graphics and VisiCalc, and you just take those metaphors. And you just watch really four decades of copy forward and we just took it and we just changed the environment. And we went from green screens to Dos, to Windows, to MacOS, to the web, to the mobile phone, but all the same core metaphors are the same.

 

The operating systems are unrecognizable from that period to now things like web browsers didn’t exist. Databases that we thought were very fundamental are completely different. The search engines didn’t exist. And yet this thing that we stare at, we, the first thing we opened in the morning, the first thing we put our new ideas and the thing that runs our board meetings, the thing that runs our town runs our compensation.

That thing hasn’t changed in 40 years, that seemed crazy to us. So when we started, we took these two observations. So the world runs on docs on apps. Those haven’t changed in 40 years. Why don’t we start from scratch? And so we built a new doc and that’s what became Coda.

Soma: [00:14:21] That’s awesome, Shishir. I want you to go back to the early days of Coda. You started the company, you brought on the first set of people, and the founding team for Coda.

And you started working on what I call the first MVP, the first version that he wanted to put out and see what customers thought about it kind of thing. When you did that, what are the feedback like? Did it catch on like wildfire? I would love to hear the journey that you went through to a place where you found initial product market fit.

Shishir Mehrotra: [00:14:51] Yeah, I think the, I always love this question for entrepreneurs. Cause I feel like it’s a debunking of real products appear. And it just seems Oh, that must have caught right away. And then you go look underneath and you see what the iteration went into it. One thing we decided when we started the company was we decided to start it in stealth, which is not a typical decision.

And there was a bunch of different reasons for it, but the main one was, I didn’t want the team to be distracted. And I felt like we had a number of prominent people in the company and. And backing the company and so on. And I thought if we spend time talking, we wouldn’t really be talking about the product.

So we basically told the company we’re not going to ship, or we’re not going to talk about the company until we can shift the product and let the product lead the story. Which I had no idea how long that would take. So we got started, at the same time, my philosophy was don’t build in a vacuum.

So as soon as possible, if you want to get people onto the product, so our first milestone was just getting to our own usability. You can sometimes we call that a dog food milestone. We had a particular use case we had in mind for that. The company was only six or seven people at the time and we basically converted our planning and task tracking system into Coda.

And then, so we’re feeling pretty good about that. We’re about four or five months in, and we say, okay, let’s, let’s find someone else. Let’s find someone other than us to do it. And so a friend of mine, a guy named Nolan Lavinsky was starting a company and there were also about six people.

And so I called them up and said, Hey, this is working well for us. Would you try it and give us some feedback? And, gladly agreed and said, I’d be happy to do it. So we had a little dashboard that tracked our daily active users and it only went from zero to six because that’s how many people they had in the company.

And then one day this thing hit zero and we wait a day and it’s still at zero the next day. And I call up Nome and I say, I said, what happened? Did you guys go on vacation? Are you having an offsite or for, he says, no, actually I’ve been meaning to call you and tell you.

We, we had a team discussion and, I have some news for you. The team all told me that if I make them keep using Coda, they’re all gonna quit. And so we had to pause and I, my first reaction was okay, I don’t know how you’re going to sugarcoat this. That sounds pretty extreme.

And, and he said, but I have some good news. Okay, what’s the good news. And he said, they’re all totally aligned on the mission for where you’re headed. They just have lots and lots of feedback on things they think you should be fixing. And we’ve built a list of 30 things that you should go work on.

And if you get these things done, we’d be happy to try again. And the interesting thing about this journey and this product, and by the way, that process repeats itself, many times we, Started the company in 2014 and we actually didn’t launch Coda 1.0 until February of 2019. About four and a half years after.

And so it was a much harder product to build than I expected, early on. And I think part of the reason, and we would hear that pattern of feedback over and over again. I totally believe in the mission. I totally understand where you’re headed. I love the promise. Can you fix these 30 things?

And gradually that 30 would go to 25 and go to 20 and so on and you get there, but everybody’s listed 30 was a little bit different. So it wasn’t just like, you could just keep working the same list and I think one of the things about building a product like this, I was talking to a friend of mine who’s deep into the video game space and builds lots of a video game.

And he says, there’s two types of video games you make, there’s some video games where you make a one level and you put it out and you don’t even bother making level two. You see how a level one goes. And once people start beating that, then you make level two. And then he was telling me this other game they made, that was this big Star Wars game.

And they worked on it for, five years. And no one of the components actually work together at all until three months before launch. And the whole thing made no sense until it all worked together. And he said, there’s a sort of two different types of products he sometimes built. And I didn’t know it at the time, but that’s what Coda ended up being.

And I think the reason for that is fairly simple Coda is a product with very high ambitions and aspirations. It’s an empowerment product. but it’s also displacing a set of tools that as I said, had been around for 40 or 50 years. And so the expectations are incredibly high. And so it became a part of, I sometimes describe Coda, like a piece of music when, when one note is off, the whole thing just sounds wrong.

And so you’re constantly finding all those different areas. So anyway, the process of building Coda was very deep interaction with customers and lots of love your vision. Fix these 30 things. I think it worked.

Soma: [00:18:59] Got it. That’s fantastic. Because like you said, most people looking at it from the outside think Hey, you go build something and then boom, it takes off kind of thing.

And maybe occasionally it does, but for a lot of people who now put something out, listen to customers. Iterate and then go through the process in a tight loop fashion, and sooner or later you get to the right place kind of thing. So that’s great to hear your sort of story during the early days of Coda.

Shishir Mehrotra: [00:19:22] Very few people know this, but YouTube started as a dating site and YouTube was more of an overnight success than most, but even there, the migration was, was meaningful.

 

Elisa La Cava: [00:19:34] I loved hearing, about how, the feedback from your early customers, giving you 30 different list of 30 things.

And then today, what you called a meta site, you can create to do lists, brainstorm ideas, manage projects, publish websites, the capabilities are incredibly powerful, and growing and endless.

But on the fun side of things. What have you seen? What’s a neat and perhaps unusual or overlooked use case you’ve seen a maker or user use Coda for so far?

Shishir Mehrotra: [00:20:06] One of the fun parts about working on platforms is that you’re constantly surprised at what people do and YouTube was similar.

I’d walked into YouTube some days and you’d look at it and you’d say, I can’t believe people did that sometimes in a good way, sometimes in a not very good way. And I think Coda has a similar element to that. There’s an incredibly long tail of what people do. Pick one fun example. There’s a venture firm called Madrona that’s apparently making investment decisions in Coda, which I think is actually a really fun one and maybe joking aside is, I think, is a really good representation of how to think about very fundamental processes a little bit differently. And I think that one is a great example of removing bias and a really hard process, and avoiding group think and really soliciting and getting the most out of a partnership, which I think is a really hard thing to do.

Let’s see. Other interesting use cases. One that sort of outside of the traditional teams using Coda to run themselves, Sal Khan’s building one right now called schoolhouse.world, which I think is, I think it was really cool. And this one was interesting, Sal is actually an old college buddy of mine.

He and I both went to school together at MIT and known each other for years, both ended up marrying our college sweethearts and actually lived just a couple of miles from each other. We launched a feature in Coda called publishing where you can publish a code of doc as a website.

And, when we launched that, he emailed me and said, Hey, could I use this to build this thing I’ve been meaning to build? And apparently his basic idea is that Khan Academy is his primary creation, which is, most people know of as a great educational site. He wrote a book called One World Schoolhouse.

And, at that time, he bought a domain called schoolhouse.world, which is, the, his sort of working view is that the boundaries of what we consider school to be will shift from being physical, to being encompassing of the whole world. And so the way the site works is pretty simple. It’s a doc where anybody can sign up as either as a tutor or as a student, and describe what you want to get to and you get match made to different group tutoring sessions.

And, it’s really interesting. It’s being run by a group of volunteers. And I think one of the, one of the really interesting things about it is Sal called me and asked about this and it was up and running in a weekend, because it was so easy to make. So I think that is a really interesting one, I think the breadth of use cases is really fun. It’s really inspiring. It’s really challenging. Building a product that can actually handle all those use cases is not easy. And you can imagine everybody’s list of 30 different things to change is very different across that spectrum, but a lot of fun.

 

Soma: [00:22:33] Awesome, Shishir. Particularly, now I do want to make a plug into Seattle here. I know that earlier on you decided that Hey, as you think about creating a distributed team, that one of the locations you are going to build a team around is in the greater Seattle area in Bellevue.

And given the amount of technical talent, particularly, but in general, the technology ecosystem talent that’s available here, I’m glad that you made that decision earlier on and hope you are happy with the decision so far.

Shishir Mehrotra: [00:23:00] Oh, the first, I think the second person we hired, it was Nigel Ellis who was running engineering for SQL Server at the time, on my old teams.

And, we had this sort of debate about it and said, are we, I know to show you talk about distributed teams being better and so on, but are we really ready to do this and discussion with our board. And honestly, most people’s reaction was that’s a little bit nuts. Like you’re six people. Like, why would you want to be split in multiple offices now?

And you all live near each other. That seems crazy. And one of the arguments I made, I think that distributed teams work better. And I also think that teams that start distributed have a much easier time staying distributed.

Actually one of my pet peeves is when people use the term remote. Remote I view as a pejorative term, a remote implies a headquarters. And I think if you think that way, if you think headquarters and remote, you’ll build one culture. If you think distributed, you have built a very different culture.

And so we started with that and it was an easy case to make it was, Nigel was great and he’s like a great person to hire. And by the way, there’s like thousands of other great engineers in Seattle that are clearly qualified to work on Coda and will be very relevant to us. Why would you box them out of being part of our journey?

And that turned out great. And we’ve got a great, thriving team in Bellevue and now all over the country and all over the world. But I think it was very helpful and setting the right scaffolding for building a distributed team.

Soma: [00:24:19] That’s great. Hey, Shishir in building on the culture that you talked about a lot so far, there is one other thing that I’ve heard about Coda, both from you, as well as from other people in the ecosystem that I want to start off and ask you about. In the six years that you’ve been around you.

You have a tremendous track record of what I call close rates of candidates. Particularly Hey, when you make an offer to a candidate, I hope many of them, they can actually end up joining. You have a very high number related to pretty much any other startup that I’ve encountered in the last many years.

Tell me what makes you and Coda and the team so special that you have such a high rate, I guess I have heard through the grape vine, it might be related to how you take care of the employees. It could be a charming personality. It could be a vision, maybe your equity policy, maybe all of the above.

So I’d love to hear, what is the reason for the success. And I think this is something that every entrepreneur should pay attention to.

Shishir Mehrotra: [00:25:15] I’m going to give you a right brain and left brain answer to this question. And thanks for the positive thoughts. I think we do well.

I’m sure we can do better. When I was debating, leaving Google to start Coda, as mentioning I was going through my two questions. So I have an idea I can’t imagine not working on and a person I can’t imagine not working with. And, I’m gradually getting conviction on both things and the idea I just couldn’t stop thinking about.

And Alex and I were very clearly like the right pair to go work on this idea. But I was still pretty resistant to starting a company. And a lot of it was because I had a friend of mine who had started a company and I was talking to him and I said, Hey Alex, so I’ve been talking about this idea.

I can’t stop thinking about it. I think I should. I think I should start this company. And this friend of mine said, she said, you can’t do that. And I said, why not? And, and he said, there’s a thousand reasons, but she should, let me give you just one reason. He said, what’s your, close rate for hiring people into YouTube.

And I just come out of a meeting with my HR lead and, and so I had the stats like right on my tip of my tongue. It was about 92%. Like it was. And you to remember at this time, YouTube was like a great place to work. Is startup inside big company, like big mission? well known product, lots of scale, but lots of opportunity for innovation.

We were pretty good at recruiting people and it was very rare that we gave offers and people didn’t accept. And so I’m talking to this friend of mine, he says, okay, that’s interesting, 92%. We sit in the forties. And we said, I spent all day long trying to find the people that are too tall for Google and too fat for Facebook.

and that was his analogy that I’m quoting him, not me and it struck me, he said, you’re just going to find yourself trying to recruit people, and you’re not going to be able to recruit the best, and you’re going to drive yourself nuts.

And for me, that sounded terrible. Like I really wanted to work on this problem. And I thought I had someone great to work with, one of the things you get used to working in a place like Google, Microsoft, and so on is you work with great people and, people that are really talented.

And the idea that I’m going to go try to find second servings from each of these companies, that sounded really terrible. And so I went and had this conversation with reading him with Reid Hoffman, and Hamilton, which ended up being the primary financier is of Coda.

And I talked to both of them about it and they both told me, look, that’s not, what’s going to happen to you and we’re going to help you understand why. And I think those conversations were really critical and me deciding to start Coda. And, so right brain, left brain. The right brain side of this is people join missions.

And the, if you have a big, bold mission and you can get people excited about it, people will find that same level of enthusiasm that you feel in going after this mission. And if you look at people joining Coda, many of them, when they describe why they’re joining and so on, they’ll describe a lot of left brain things.

I’ll talk about in a moment. But they’ll all start with I just thought it was a chance to build a thing that really mattered. And that’s a thing that when you get those opportunities, you get excited about and you really feel motivated about, and that can drive a lot. And I think it’s one of the things I ended up coaching entrepreneurs on a lot is how to tell your story in a way that lets people go on that journey with you.

And there’s probably a version of it that caused you as an entrepreneur to feel that convicted about it. But sometimes telling that and helping people feel part of it is really important. And I think we do a pretty good job with that. I think Coda has a big mission and has it has a good chance of impact lots of the world and lots of different aspects of the world.

And yeah, somebody, if I just talk about, for example, we talked about with distributed teams and changing how you think about bias and that’s the type of thing that would not be obvious. I just told you, Hey, we’re going to go rebuild office because we think it hasn’t been rebuilt in 40 years you probably would be excited, but maybe not that excited, but if I told you, I think the world is full of cases where whole groups of people are ignored or don’t realize their voice, their potential.

So you might be inspired by it in a totally different way. and so I think getting good at telling that story is really important. And I think, I think my, our recruiting team, each of our leaders are all very good at this and they will tell some version of this in a good way.

And it’s infectious. Then each person that turns around and tells it to the next person and so on. So I think that’s really important. I think people don’t spend enough time on this. Kenny Mendez who runs, people in operations for us is always one of the best storytellers I know. And, and has been a really good at not only doing this himself, but building a team of people that can do this well, that’s the right brain side, the left brain side, talking to him and Reid about this.

They said, look, you’re going to tell a great story. you have a pretty good network to build off of, and people want to join places where they can join other great people. And I think that’s really important, but there’s a practical side of joining a startup. And one of the things that I think companies don’t do.

And so this will get to the mechanical part of this is they’re unrealistic about the decision facing an employee and these employees, anybody you want to hire has many offers and that’s good. Like liquidity is good. The job market is generally healthy. You don’t generally want the person that is not able to get other offers.

In fact, we often offer genuinely to help people with it. I’ll help people connect with other companies. I’ll help them. I’ll reference, check for them if they, if that’s helpful and someone, because I feel like when you join a place like Coda, I don’t want you to joining because we were your last resort.

I want you to joining because you understood your options and you decided this was the best one. And if it’s not that’s okay. And I think being clear on that is helpful, but at that point, they’re going to have a decision to make, and that decision is likely not theirs alone.

Like they may have a spouse or partner. They may have a parent or family members that are coaching them, they may have an advisor or so on, and those people are gonna start left brain. And I think that from that perspective, the main thing that we end up talking about is we treat employees, making a decision to join a company as being investors.

And that philosophy is the way I think about it is I always tell people, look, when you’re joining a company you’re investing and you’re investing with your time, not with your money, but, boy, time is a way more precious resource the money. So you need to think about it in that way.

And we’ll do a lot of work too try to make this process clear to people. And, I’ll just to give a few of them first off, we’re very generous with equity. We make it such that, and that starts by, we didn’t sell that much to investors so you can get more to employees. My view is most companies end up being held too much by investors and founders and not enough by employees and it’s not a good thing.

So being generous with equity is really important. How you present the offers. I can’t tell you how many people present offers and here’s your number of shares and they don’t tell you basic information. What’s the total float of the company? What was the last round? What were the terms?

What are the gotchas in the around? Are there any special provisions? Like all these things just don’t give enough information for the person to think like an investor. And so we built an offer model that helps people run through this process. One of the things that model does is it gives an unexpected value calculator.

Which is another thing that, most employers I’ll have many employees look at equity and they’ll think of it is worth either zero or worth of a jillion dollars. The employees have no way to gauge anything in between. But smart investors know that’s not how you should think about equity.

And there’s four new investor. In a company, there’s some percentage chance of the mega outcome. And there’s some percentage chance of the mediocre outcome. And there’s some percentage chance of the zero. And yeah. And you define your scenarios and we don’t fill in anybody’s numbers and you should make your own decisions, but we just help people through that decision, give you enough information and to be able to have this conversation with your partner, with your spouse, with your family members and so on.

And, and realize that you may be excited because you want to change the world and you want to change how things operate and so on, but they want to make sure that you’re making an economically sound decision for your, for your family as well. There’s a number of other things we do there. We do a thing called founders preferred stock, which is a special tier of stock that converts a little bit closer to a preferred stock.

The ways we do the actual mechanics of the offer is and options and so on is a little bit different, but the basic philosophy is how people have that same founder level of conviction on your mission, and then treat them like investors as they make a decision to invest their time into your company.

And I thought those were like those two things together lead to building a company full of great people.

Soma: [00:33:11] We’ve had these conversations over the last couple of years, but every time sort of hearing from you about your journey about Coda journey, it’s always been fun.

It’s great. So thank you. Thank you for sharing your sort of thoughts and perspectives and your journey with us.

 

Shishir Mehrotra: [00:33:25] All right. Thank you. This was a lot of fun, lots of great questions and a great exploration. Thank you.

Erika: [00:33:32] Thanks for joining us for Founded and Funded. That was a great conversation with Shishir and there is more to come. We have another podcast coming later this week. That really goes a little bit more in depth into the future of work with Shishir.

Erika: [00:33:48]

Please stay tuned for that podcast coming up later this week. And send us any feedback that you have about the podcast. You can send it directly to me. It’s [email protected] and that’s E R I K A.

At madrona.com. Thanks and we hope you have a great week

 

Investing in Intelligent Applications – 2020 and beyond

We have been writing about the Intelligent App Application Stack for over five years – which means we’ve been investing it in for even longer. And that has not changed in 2020. We continue to believe that applications will be intelligent. What has changed over that time is both the rate of adoption and the availability of underlying infrastructure and technology to support these applications. In this post we detail four areas that we are continuing to see innovation around intelligent applications.

We define application intelligence as the process of using machine learning to create apps that use historical and real-time data to build a continuous learning system and make predictions and deliver rich, adaptive, personalized experiences for users. Intelligent applications typically have a modern user experience; a cloud-native, microservices architecture, and integrations to other systems and cloud services. We believe that every successful new application built today and, in the future, will be an intelligent application.

What benefits do intelligent applications deliver? We believe that next generation intelligent apps will allow us to:

  • Create custom workflows to automate any process
  • Process data in real-time across multiple systems of records to deliver insights and predictive capabilities
  • Build digital-first go to market and customer service models
  • Provide better services with lower delivery and customer service costs
  • Become the new systems of intelligence on top of legacy transactional systems

What are the types of intelligent applications that we are seeking to invest in at Madrona? Here are four broad categories where we see immense opportunities:

1. Automation

Many of the most impactful “intelligent apps” today focus on identifying repetitive, time-consuming processes and creating new ways to handle these workflows in a way that allows customers to focus more of their time on high-value synthesis and cognitive work. This is a cornerstone of the digital transformation that every enterprise around the world is currently going through.

The largest companies in this space today are the robotic process automation (RPA) vendors, such as Madrona portfolio company, UiPath. These companies have built horizontal platforms that allows companies to automate individual steps of a workflow, such as opening up a PDF document, extracting key data, entering that data into another system, and combining these steps into an automated workflow.

Despite the success of RPA, it is only scratching the surface of what is possible with AI and automation. With innovations in computer vision, natural language understanding, and other machine learning techniques, we are also seeing more and more companies build “RPA”-like automation into new products to create end-to-end workflows for specific use cases in industries such as legal services, financial services, healthcare, and real estate.

Some of the most interesting companies in this space go beyond automating one workflow to automating multiple workflows and creating a new integrated workflow. For example, a company like Zeitworks uses machine learning to map out a customer’s workflows in order to help understand which processes can be automated and to track how they perform over time. Madrona funded Zeitworks’ seed round in June of 2020 recognizing both the need for discovery of workflows prior to applying automation and the fact that automation for small to medium sized firms is particularly needed as the workforce and resources they need are not co-located any longer.

2. Next Generation Business Applications

Many of today’s key business systems for finance, HR, sales, and customer support were built decades ago, with software architectures that have not changed for the last twenty years. While these companies have built large businesses around certain types of customer behavior, they are often unable to innovate at the same pace that modern companies need.

We believe the most successful next-gen business applications will compete with their legacy alternatives by attacking a small portion of what their legacy competitors offer today or completely reimagining a business process that can only be enabled with modern software architectures.

For example, in the travel and expense space, Concur was founded in 1993 and built a massive business digitizing a manual process where paper receipts and expenses were passed from employees, to managers, to FP&A teams. Modern startups are transforming this process by reorienting around purchase data instead of receipts and forms. Rather than waiting for a month after a purchase is made, modern tools like Center deeply integrate credit cards with enterprise grade software to process expenses as they occur to give managers real-time insights into employee purchasing behavior and budgets.

HighSpot, a Madrona portfolio company, is another example of an intelligent application that uses integrations and data from multiple systems to help sales teams find the right content and relevant guidance for each conversation. By using data from CRM systems, email, and other workflow tools, their system is able to score content and understand what engages customers and drives revenue.

These types of workflows and systems are possible today because of modern microservices architectures that can process data in real-time, stream data to and from other systems, and convert data and insights into immediate actions. While many of these modern platforms start with a small feature like better insights, better UI, or better data, we believe they have the potential to eventually replace legacy systems.

3. “Avant Garde” Applications

Photo of an Amazon Go cashierless store.

“Avant garde” applications create completely new experiences and products by using machine learning – services that just weren’t possible before the combination of low-cost cloud computing, massive amounts of data, and new machine learning algorithms.

ML breakthroughs in fields like robotics and computer vision have created self-driving automobiles, which enable completely new vehicle form factors, business models, and services. Alexa, Siri, and Google Home’s voice assistants enable new interaction models that would not have been possible without advancements in natural language processing.

Many of the companies in this category are pioneers in bringing important new technologies such as computer vision, deep learning, robotics, and NLP to market, so it is a very dynamic space to watch because it sits at the intersection of massive markets, cutting-edge technologies, and novel business models.

For example, Amazon Go has created a completely new shopping experience by using computer vision technology to reimagine the shopping workflow. This allows for the construction of stores with new layouts that don’t require cash registers at the exit and may one day allow for retail stores to adopt new business models as well.

4. Intersection of Innovation spanning Life Science and Data Science

This is a vertical specific intelligent application category. However, given the potential opportunity size and impact, we have called it out seperately.

Whether it is in the field of diagnostics, therapeutics, or healthcare operational efficiencies, the availability of massive data sets combined with applied ML/AI is revolutionizing what is possible in the fields of life science and healthcare. For example, a company such as Adaptive Biotechnologies leverages decades of research on the immune system, next generation sequencing, and machine learning in order to detect changes in the immune system to diagnose disease.

While these companies can become massive winners, they may also be harder to measure and monetize in the short term. However, as early stage investors, we are excited to continue exploring investments in this category.

Over the next decade, we believe that every successful new application will be an intelligent application, and this will lead to many opportunities to build enduring software companies. If you are working on building an intelligent app in one of these categories, we would love to meet you and learn more! Our contact info is linked in our byline!

Investing in Low Code No Code Development Platforms

This is the third in our series describing our investment themes for 2020 and beyond.

Software is eating the world, but today, most organizations have a limit on how quickly they can build software. While teams want to automate processes, get insights from their data, and build better solutions to problems, there are simply not enough people who can write software to do these things.

At Madrona, we believe low code and no code development platforms unlock new opportunities for people and companies around the world to accelerate the promise of “software eating the world” by (1) helping experts work more effectively and (2) giving more people the ability to write software.

(click into the image for full screen)

We think there are three major trends driving the adoption of these low code and no code platforms:

  1. Companies want to move faster, increase productivity, and reduce costs.

In today’s business environment, companies are shifting traditional ‘development’ work from IT and engineering teams into line of business teams. These teams want platforms that allow them to do their jobs without requiring engineering cycles.

  1. Cloud-native apps make it cheaper and easier to build

New application delivery models are making it easier than ever before to build powerful, customizable software. Not only are cloud-based applications cheaper to get off the ground, they also make it easier to connect to other apps, services, and databases.

  1. There are more “makers” than ever before

There is a generational shift underway in the technology world, and software “users” today have both higher expectations and more “coding” fluency than ever before. Users expect to leverage software as a flexible tool to suit their needs, and they are willing to spend time customizing tools for their individual use cases.

These trends are creating an inflection point in the adoption of low code and no code platforms that will drive new opportunities for makers, startups, and enterprises. We have already seen each of the major cloud providers bring low code products to market (Microsoft PowerApps, Amazon Honeycode, and Google AppSheet), and we are excited to see more innovation from startups as well.

Types of Low Code and No Code Platforms

We take a broad view of low code and no code platforms as products that (1) help experts work more effectively and/or (2) give more people the ability to write software. We take this broad view because we believe this is a massive technology trend that will affect every “horizontal” and “vertical” software market.

However, to help us think through the different opportunities, we like to categorize products as horizontal point solutions, horizontal full stack solutions, and vertical platforms (which may also be either point solutions or full stack solutions.)

The distinctions between these categories are very blurry, and many point solutions evolve into full stack solutions over time, and many full stack solutions are often better at one piece of the application development lifecycle. But these categories help us understand the types of customer pain points and workflows that each product is trying to solve.

Horizontal Platforms

We like to think of the “application stack” for low code and no code platforms in three parts:

  • The Interface Layer – The interface layer is the end-user facing part of a low code application that is typically used to either input data (e.g., better forms or data collection tools) or visualize data and outputs (through charts, apps, etc.)
  • The Data Flow Layer – The data flow layer handles the basic create, read, update, and delete operations to the database in addition to handling business logic and rules and connecting to third party services and data
  • The “Database” Layer – The “database” for a low code application can range from a Google sheet to a MongoDB Atlas instance. We think of the “database” for a low code app as the key system of record that the interface and data flow layers are interacting with

(click image for full screen)

When we take a closer look at a market map of horizontal low code and no code platforms, we can see that the lines between different types of solutions is very blurry. Companies operate between multiple layers of the stack, and while we see some distinctions between companies that focus on internal applications vs. external applications, the lines are very fluid.

Some of the key opportunities we see for horizontal companies are:

  • Automate workflows that sit between engineering teams and other functional or business teams – Tableau is one of the best examples of this type of opportunity. Business teams want to have real-time data on performance, but the need to ask engineering teams to constantly make small changes to data queries and refresh dashboards has created an entire industry around low code solutions for business intelligence
  • Build more functionality into familiar interfaces – Coda is an interesting example of a company that uses “familiar” interfaces like Word docs and Excel formulas to create a new type of doc that can have app-like functionality. By starting with familiar interfaces, it is easier to teach users how to “code” in a low code or no code setting
  • Build horizontal products with a vertical GTM focus – Unqork offers low code solutions for forms and workflows that connect to legacy databases. Their initial GTM approach was to focus on selling to financial services firms, but they were able to keep their product flexible enough to work in other industries such as government, real estate, and education

Vertical Platforms

Historically, many of the most successful low code platforms started with a very narrow use case, and as they became more powerful, they expanded beyond what their original intended use. For example, Epic Games’ Unreal Engine was originally built as the game engine for a first party game. Eventually, the team decided it could be more valuable as a game engine for other game developers, and now it is often used as a tool to for movie visual effects and AR/VR app development.

We are very interested in these purpose-built platforms because they typically start with a specific customer pain point in mind, and this helps drive focus in the early days of getting a company off the ground. Over time, as markets grow and new opportunities arise, they also create the potential for game-changing companies – like Epic or Unity.

Some of the verticals that we think are particularly interesting right now for low code and no code applications are ecommerce, data science and machine learning, and infrastructure. We picked one industry vertical in ecommerce to illustrate vertical specific solutions, but we see that trend in other verticals as well.

Ecommerce

Within the ecommerce vertical, there are several opportunities for new startups to create new products based on the broad ecommerce value chain components. The broadest opportunity – where Shopify is already the clear leader – is a service to create end-to-end ecommerce stores and sites from scratch.

Shopify is the leader in this segment today, but we think there are several parts of the market that are underserved by Shopify’s current product. For example, we would love to see a company offer a seamless omnichannel commerce solution powered by a low code backend. We also think there will be companies that can build large businesses around category-specific services like frozen fulfillment, CPG, subscription, or alcohol.

For other pieces of the value chain, we see opportunities for companies to create better consumer-focused experiences, deploy custom functionality such as chat or checkout in an existing site, and offering better analytics and planning tools to existing stores and platforms.

Data Science and Machine Learning

Data science and machine learning is one of the fastest growing parts of the technology world, and this is a market with a clear need for tools that make it easier for more people to take advantage of the latest developments in machine learning research. For this market, we take a customer-centric approach to thinking about the user profiles of different types of data science “practitioners.”

The first category of users is data scientists and ML engineers who are familiar with the existing tools. These users already have established tool chains and workflows, and they are more comfortable working in code, so may find low code and no code tools less flexible than code-driven tools. One key opportunity for these users is tools to assist with data wrangling and ML Ops, which are the less enjoyable parts of their work today.

The next category of users is software engineers, who are comfortable writing code, but may not have a background in data science and machine learning. For these users, there are several tools that provide an onboarding ramp to data science, but they may prefer to use “high code” tools over time.

The last category of users is non-technical business users who do not write code today and also may not understand fundamental data science concepts. For these users, we believe there is a large opportunity to give them the tools to learn more about machine learning and apply ML techniques to their work. Today, we see tools that strive to support a myriad of different analytical problems. However, we think applications that focus on specific problems such as LTV prediction, churn prediction, or fraud detection will be easier to adopt by business users. The key challenge will be integrating the machine learning services into an existing process or workflow to make insights and recommendations effective.

Infrastructure

Unlike several of the other verticals we discussed, this is a category where most of the tools are built for software developers. Some of the key customer pain points that we see companies solving in this space include helping developers quickly set up a back end “as a service” for a new application, automating DevOps workflows, and building API gateways to connect applications.

Companies like Busywork and Darklang are building “backend-as-a-service” platforms to help front end developers and full stack developers quickly set up backend systems, in order to reduce the complexity and cost of building distributed application backends. Instead of building these systems themselves, companies can use a service to get something up and running quickly. Another example of a “backend-as-a-service” offering is a product like Auth0, which allows developers to quickly build authentication and authorization into their applications.

Another common development paradigm we are seeing more and more frequently is leveraging best-of-breed services from multiple third parties in a single application. One company that helps with this is Temporal.io, which is an open source workflow engine based on the Cadence OSS project from Uber. This allows developers to focus their time on writing business code, rather than the “glue” to piece different services together.

Conclusion

Digital transformation is happening faster than ever before, and we believe one of the major drivers of the next wave of digital transformation will be the transformation of software from static databases to dynamic applications that users can customize and build themselves. Over the next decade, low code and no code systems will evolve from separate systems to becoming a core design philosophy of software systems, and we are excited to meet with companies and entrepreneurs who share this vision!

 

Madrona’s Investment Themes for 2020 and Beyond

Over these last months of quarantine, we at Madrona have remained very busy, first and foremost working with our portfolio companies to help them navigate the economic turmoil of a global pandemic. Second, but also importantly, we have continued to invest, adding eight new companies to our portfolio since quarantine began in March (Fauna, VNDLY, Go1, Zeitworks and four others still unannounced). This continued active pace of investment exemplifies both our commitment to the long-term opportunities we have identified, as well our belief that downturns can be the best time to invest.

As we have been quarantined in home offices in front of Zoom and Teams, we have also taken the opportunity to step back and revisit our investment themes and think about which trends we continue to be most excited, which are emerging, and which perhaps are being accelerated (or dampened) by the aftermath and “new normal” of COVID-19.

It has been 18 months since we last posted about the investment themes that are driving our activities at Madrona. When we took a fresh look at our current thinking on technology trends that we believe will drive the industry in the next five, ten, or 20 years, the picture that emerged shows significant consistency with our view of the world 18 months ago, but also interesting new opportunities and trends. The figure above illustrates the overall areas we find most compelling for new company and investment opportunities.

In this post we offer a preview of the themes and will follow this up with deeper dives on the areas outlined in the image above. We work as a team to fully investigate and build our investment themes and you will see many from the Madrona team as authors – please reach out to us with ideas and your thoughts!

At the center of our investment themes, we continue to see a massive opportunity for companies to create businesses around intelligent applications, fueled by machine learning and modern user interfaces. We believe that every successful application being built today should be an intelligent application, with a data strategy and continuous learning system at its core. Intelligent applications have been the single largest area of investment for us over the previous several years, and we expect this to continue for the foreseeable future. We will continue to invest in the next generation of line-of-business applications being reinvented by machine learning and cloud native delivery. Read more about the areas we are seeing opportunity in intelligent applications in our deep dive.

As the world continues to struggle through the COVID pandemic, we also see a massive acceleration in the emergence of technologies enabling the future of work. This trend had been evolving for the last several years and the current environment has created an order of magnitude acceleration, as businesses of all sizes rush to find new intelligent applications that help them collaborate more effectively when all employees are remote, build and retain more diverse and distributed workforces, and prioritize digital-first workflows and processes that have remained largely “in-person” and workplace focused. Read more about the work, the workplace, and workforce in our deep dive.

A major new focus area for Madrona is the intersection of innovation between machine learning, intelligent applications, and life science. We touched on this opportunity in our investment themes 18 months ago, and subsequently invested in several exciting companies including Twinstrand, Nautilus Biotechnology, and Terray Therapeutics. As our partner Matt wrote when we announced our recent large investment in Nautilus: “Today, these domains are coming together to transform the ways we understand and improve life and health. The biological and chemical sciences are intersecting with computer and data sciences in precision medicine, digital pathology, proteomics and more. At Madrona, we believe these intersections of innovation will be at the forefront of major breakthroughs in research, analysis, diagnostics, clinical processes, preventions and cures.”

Next, the march to the cloud and broader adoption of the cloud computing model by enterprises continue to create myriad opportunities for next-generation software infrastructure companies — despite the increasing dominance of the hyperscale public cloud providers. These steady improvements to software infrastructure enable and increase the pace of innovation for all the applications higher in the stack that leverage these cloud services. Enterprise need for better usability, manageability, security, cost-savings, and performance across diverse devices, cloud platforms and environments will drive new business opportunities that provide hybrid and multi-cloud management, infrastructure automation, and new architectures that leverage serverless and event-driven architectures. Read more in our deep dive on The Remaking of Enterprise Infrastructure.

Another investment theme created by the need to move faster, increase productivity and reduce cost is in the area of low-code or no-code platforms and applications. The next generation of workers is more tech savvy, and there are more “makers” in business teams and organizations who want to build things directly and not wait for IT, engineering or the data science team. These range from developers who need to incorporate ML directly into the applications they are building to information workers who become citizen developers in order to quickly solve business problems. Read more about how we think about no-code/low-code in our deep dive.

While we have invested a somewhat higher percentage of our last several Funds in B2B companies, we continue to strongly believe in and invest in new consumer services, often where the digital and physical worlds are fused in a way that create a virtuous cycle to provide a more compelling and fully integrated experience. This digital transformation of consumer experiences, where mobile-first applications streamline, simplify, and save consumers time and money, is a core pillar in our investment themes going forward. Read our deep dive on the areas we see changing dramatically in the next five years here.

We are eager to engage with all of you in the community around these updated investment themes. Each time we have published our thoughts in the past, we have been energized and humbled with the feedback we have received – from founders whose vision hew closely to one of our themes to constructive debate on how we are too early or too late with our ideas. In the coming weeks, we will post six deeper dives into our themes around software infrastructure, intelligent applications, the future of work, the intersections of innovation, low-code/no-code platforms, and the digital transformation of consumer experiences. We can’t wait to further discuss, debate and learn from all of you. In the process, we look forward to investing and working alongside some of you to build the next generation of companies that address these exciting areas of innovation.

Send us an email or connect with us on Linked In (all contact info is in our bios which are linked above).

 

 

 

Innovation where Life Sciences and Computer Science Meet

The Pacific Northwest is a major hub of tech innovation. It is also a hub for life sciences research, biotech and healthcare innovation. The past several years have brought increasing convergence of these disciplines, most notably the nexus of life science, computer science and data science. This combination has been a driver of new breakthroughs — i.e. use of machine learning in discovery, diagnostics and therapeutics.

Our region is home to two of the top market cap companies, Amazon and Microsoft, who are both leaders in cloud technologies. These companies are defining and building the scale infrastructure and platforms, including major advancements in Machine Learning (ML) and Artificial Intelligence (AI), for next generation applications. Major research institutions such as the Fred Hutchinson Cancer Research Center, Allen Institute for AI (Ai2), Allen Institute for Brain Science and a growing ecosystem of companies (e.g., Adaptive) are starting to leverage the power of data, algorithms and computing power to develop breakthrough research and products driving critical improvements in healthcare and global health.

The convergence is enabling new opportunities in the broader healthcare and life sciences markets – spanning from traditional healthcare IT to digital health to diagnostics to next generation therapeutics and automated scientific discovery. We have already invested in several companies in this area – Saykara which is bringing NLP and AI to the world of medical scribes, Accolade which helps employees get the most out of their healthcare plan using software intelligence and people, and Envisagenics, the recipient of the Madrona/Microsoft AI prize which is applying AI and high-performance computing to uncover novel cures in RNA sequencing data.

In working with entrepreneurs and the local industry, we’ve looked at the broad market, divided it into “more healthcare” and “more life sciences” and identified areas of specific interest where we see substantial opportunity for software and data-science/AI driven innovation and are within our expertise. Our map of this intersection is below and we will highlight a couple areas of particular interest.

Diagnostics: In the area of diagnostics, ML and AI techniques are already empowering next generation clinical decision support services into the market. The application of computer vision to radiology and pathology is one example. Companies such as Zebra, Viz.AI, Imagen, and others have had AI/ML based medical diagnostics achieve regulatory approval in areas such as stroke diagnosis, atrial fibrillation detection, fracture diagnosis, and others. In the area of cancer diagnosis, new companies such as PAIGE and PathAI are making major strides. In the past year, we’ve seen an increase in new AI-powered offerings achieving regulatory approval in a broad range of diagnostics from stroke to wrist fracture to heart & lung related diagnostics and others.

Infrastructure: To support research and development of new drugs fueled by an understanding of genomics data, there are several important infrastructure categories. One thing we’ve noted over the past year is that our software and infrastructure companies are seeing growth in this vertical. One of these is Qumulo, which provides next generation file storage for institutions like the Carnegie Institution for Science which works with terabyte-size data sets alongside millions of tiny sequencing files.

Analytics: On the more traditional IT end of things, we see an opportunity for analytics that overlay systems for running labs, processes, healthcare systems and more to provide better insights and help drive operational efficiency and improved care. KenSci is a good example of a company working on analytics for large hospital systems.

Data: And, underlying each of these categories is a significant need for data. Data is what will power diagnostic services development, drug discovery, clinical trial matching and many more clinical and research applications. There is a need and opportunity for data providers and ecosystems to leverage the data to drive the innovations we all foresee. Existing players such as Prognos, Patients Like Me, Tempus, and RDMD are all working on this space and we are excited to see the next wave of innovation in data acquisition and management.

As 2019 unfolds we will continue to share our thoughts and if these areas are of interest to you, please engage us.

The Future of Retail – a 2019 Investment Theme

This is the fourth and last deep dive into the technology investment themes we outlined earlier this year that we will delve into in 2019 and beyond.

Twenty-five years after the launch of Amazon.com, e-commerce represents about 10% of all retail sales. From zero to ten percent of a multi-trillion dollar sector in just 25 years is astronomical growth, but as convenient as shopping on the internet has become, it’s notable that nine out of ten transactions still take place offline. And yet, as we marvel at the pace of change and innovation in e-commerce over the last two and a half decades, the technological advancements in the physical retail experience are far less notable. In Seattle, we are at the center of much of this change and innovation, and while e-commerce and digital marketplaces remain a core investment theme at Madrona, we see a tidal wave of change coming to the physical retail world, led by technology, and are excited about several trends that will shape the future of physical retail.

“As we marvel at the pace of change and innovation in e-commerce over the last two and a half decades, the technological advancements in the physical retail experience are far less notable.”

Retail Infrastructure as a Service

For digitally native brands that can open an e-commerce store by launching a website or pushing an app to iTunes, opening a retail store feels like a trip to the stone ages. Hiring brokers, committing to long-term leases, spending hundreds of thousands of dollars on build-outs and inventory, hiring and training teams, and driving local awareness . . . a lot goes into it.

Indochino, a Madrona portfolio company and digitally native brand that sells custom clothing, now has 40 retail stores in North America. These stores generate a significant portion of the company’s revenue and are its best customer acquisition vehicle. New customers who engage with the brand through Indochino’s retail stores have a better first-time experience – they get fitted for their custom garments by a professional, they can touch and feel the fabrics, and they get guidance and advice to help them make their custom suits, shirts, and khakis truly their own. Indochino’s online business is growing fastest in markets where it has physical retail presence. But as successful as Indochino’s retail strategy has been, the process of finding, securing, and opening new stores is a grind. And opening dedicated stores like Indochino won’t be the right answer for all brands.

“We believe there is an opportunity to create retail ‘infrastructure as a service’ that is loosely analogous to the cloud IaaS we are all familiar with today.”

We believe there is an opportunity to create retail ‘infrastructure as a service’ that is loosely analogous to the cloud IaaS we are all familiar with today. Like their digital counterparts, retail IaaS providers will offer a mix of pure infrastructure, paid for on a usage basis (without long-term commitments), and value-added applications on top of the infrastructure.

For example, when establishing a physical retail presence in a market, a brand will be able to rent a whole retail space, open a shop-in-shop, or lease shelf space next to related brands in a store. The IaaS provider will take care of fixturing, IT, point of sale systems, and other infrastructure necessary to support the brand’s retail business, all built into the monthly price. Brands will be able to staff stores and/or sections with their own dedicated employees (hired and trained by the IaaS platform provider) or can rent partial time from the IaaS’s shared labor.

The platform provider will provide a range of value-added services back to brands (available a la carte), including store and shopper analytics, sales and inventory forecasting returns and exchange management (including for the brands’ online orders), and a host of local promotional services.

As far out there as this might sound, there are already a number of early-stage companies that have launched to address aspects of this opportunity, including Leap, FourPost, Showfields, and b8ta. We are following these companies and keeping our eyes open for new entrants that are meeting the needs of digitally native brands.

Digitizing the In-Store Experience

Nearly every aspect of the in-store retail experience can be improved by technology, whether applied in a clean-sheet (new retail format) context or retro-fitted onto existing retail stores. We are actively looking at and interested in opportunities across a few key areas of the retail experience, including:

  • Frictionless Checkout Nobody likes to stand in line at the end of their shopping experience to pay. The magic of frictionless checkout at Amazon Go, Bingobox in China, and other similar concepts, is pushing retail incumbents to raise the bar on the payments experience in their retail locations. Startups like Zippin, Standard Cognition, AVA Retail and larger companies alike are racing to bring sensor fusion, machine learning, and human-in-the-loop reviews to quickly and accurately identify the items you’ve selected while shopping and automatically charge you for them within minutes of leaving the store. Other approaches to frictionless checkout (e.g., mobile, scanner-driven self-checkout) will get earlier traction in big box retail environments but are less defensible and represent a smaller long-term opportunity.
  • Inventory Awareness and Prediction As retail stores become micro-warehouses and items on store shelves can be picked by the end consumer, by third-party delivery agents, and by the retailer’s own employees (for in-store pickup, online order fulfillment, or delivery), understanding quantity and state of inventory in a store becomes more complicated and more important. Systems have been developed to enable e-commerce fulfillment centers to track real-time location and state of inventory; something similar and yet more sophisticated must be developed for physical retail.
  • Merchandising (compliance) The growth of Amazon’s multi-billion dollar advertising business is a reminder that brands spend massive marketing budgets to convert sell-in to sell-through and deploy a large chunk of these budgets to promotions at retail. Premium shelf placement, endcaps, and other marketing vehicles drive consumer awareness and demand; brands pay for placement, but at the store level, they don’t know whether their products are merchandised appropriately or have sufficient stock on the floor. There are some relatively mature companies addressing this in a manual way and some early-stage companies experimenting with robotics and automation to enable scale and precision. The size of the prize for vendors in this market remains an open question, but the potential value of instrumenting retail stores to provide the real-time visibility we take for granted on the internet is quite large.
  • Customer Analytics The insights we capture about an online shopper (what channel drove them to the site, the products they viewed, the searches they executed, the items they added to cart, etc.) mostly don’t exist in physical retail. This is another category in which sensor fusion and machine learning could bring the offline experience to parity with its online counterpart. While some journalists and analysts have expressed concern about privacy in the context of ubiquitous camera coverage in retail stores, what we can learn about individual customer behavior at retail is already well-understood about those customers online. The frictionless checkout technology providers may be best positioned to provide the customer analytics, but there may also be opportunities for analytics companies to bolt onto the sensor infrastructure to provide deeper insights.

Faster Omni-Channel Delivery

The bright line between digital and physical commerce is blurring in large part due to innovations in supply chain and logistics. As physical retail stores perform double-duty as micro-fulfillment centers, and as Amazon and its scale competitors build ‘forward-deployed’ warehouses in major cities, same- and next-day delivery will be commonplace for orders placed through mobile, voice and in-person interfaces. This will require all omni-channel retail players to leverage technology and infrastructure to enable the near-instant gratification their customers will demand while doing so in a cost-effective manner. Technology areas we are actively tracking and/or pursuing that will be enablers of this future include:

  • Virtual warehouse and fulfillment networks Companies like Flexe and Deliverr are building national networks of 3PLs with an intelligence and orchestration layer on top, enabling brands to store inventory near pockets of demand, enabling fast, cost-effective delivery SLAs that can match Amazon Prime (2-day delivery) in the near-term and same-day as these companies build out their networks and customer density. There are a handful of other companies building out their own 3PL infrastructure (versus serving as an orchestration layer on top), which we believe will enable tighter end-to-end control, but may prove prohibitive to scale.
  • Last-mile (autonomous) robotic delivery At least half a dozen venture-backed companies (marble, nuro, starship technologies, Robby technologies, Dispatch, and unsupervised.ai) are piloting last-mile, robotic delivery solutions with local food and grocery delivery services. Amazon announced its own robotic delivery program, Scout, and both Wal Mart and FedEx are experimenting in this space. With last mile accounting for >50% of total fulfillment cost, mostly derived from fuel and labor, autonomous, ground-based robots could drastically improve the cost-effectiveness of same-day delivery.

“The bright line between digital and physical commerce is blurring in large part due to innovations in supply chain and logistics.”

Clean-Sheet Retail Concepts

In China, we’ve seen a spate of new retail concepts, from Bingobox, the unmanned convenience store, to Hema, Alibaba’s grocery store of the future, to Luckin, the mobile-first, delivery-oriented coffee chain. These concepts, built on top of the payments/messaging platforms of Alibaba and Tencent (and often leveraging their balance sheets), are growing at a pace unheard of in the U.S. They are leveraging many of the technologies discussed above and others to reimagine the retail customer experience.

Why don’t we see a similar pace of innovation and application of technology to reimagine the retail concept in the U.S.? Arguably, Amazon is leading the way with its Go stores (and to an extent the Bookstores and 4-Star); and we have seen some innovation from startups, mostly in the food/restaurant space, leveraging robotics (e.g., Zume Pizza, Spyce). But we have not seen the breadth and pace of retail innovation like we have in China. While capital-intensive, we believe these types of opportunities will exist in a number of retail categories, and we are starting to explore some early concepts with former executives who helped build the current generation of retail giants like Starbucks, Nordstrom, and REI, and who have made Seattle their home.

If you are working on one of these or related areas, and would like to trade thoughts and ideas and enrich the conversation, and/or if you’re looking for a thought and capital partner to help build your business, we’d love to hear from you; especially if you want to build it in Seattle, home to the most innovative online and omni-channel commerce companies in the world.

Why and How Intelligent Applications Continue to Drive Our Investing

Intelligent Applications have been and continue to be a focus of our investing. These apps sit on top of the infrastructure a company chooses, the data they collect and curate, the machine learning they apply to the data and the continuous learning system they build. In this deep dive we talk about why intelligent applications are a central component to our investing themes and where we see the opportunities for company creation and building.

Intelligent apps are applications that use data and machine learning to create a continuous learning system that delivers rich, adaptive, and personalized experiences for users. These intelligent apps range from “net-new” apps like those powering autonomous vehicles and automated retail stores to existing apps that are enhanced with intelligence, such as lead scoring in a CRM app or content recommendations in a media app.

Intelligent apps will have a massive impact on the way we work, live, and play, and we have already been blown away by the potential in what we are seeing companies build today. Some of the most exciting intelligent apps we have seen do at least one of the following:

Enable completely new behaviors

Some of the most impressive demonstrations of machine learning are those that use AI to create new business processes and markets that completely change the way people do things. One high-profile example is Amazon Go stores using computer vision to completely change the supermarket or convenience store experience by removing the checkout process.

Another great example is Textio. Textio offers an AI-powered ‘augmented writing platform’ which draws on massive amounts of historical data to help companies write better job descriptions that will attract higher quality applicants. Both of these examples use AI to create new processes that result in better experiences and better outcomes for their users.

Drive 10x (or better) process improvements

AI automation and insights can also be used to optimize existing processes and workflows. Automation using AI is at the cornerstone of what every enterprise is going through in terms of digital transformation. For example, UiPath’s RPA platform allows companies to drastically reduce costs by automating a wide variety of software based tasks using UiPath’s “robots.” While the UiPath platform is early in its journey to becoming an intelligent app, it is already helping its customer drive 10x process improvements.

Suplari also uses AI to improve existing business processes, namely to analyze purchase behaviors to better understand how to drive cost reductions and manage supplier risks. While Suplari’s customers may have individual processes to reduce software costs through deduplication or to identify opportunities for savings in contract renewals, using AI to proactively identify the best opportunities allows their customers to realize large efficiencies in their procurement processes.

Integrate silos (data and workflows) and capture value

Another great opportunity for AI companies is to combine data and processes to allow companies to combine different parts of the value chain and capture more value. For example, Affirm uses machine learning to approve consumer loans and uses these loans to help ecommerce companies improve shopping cart conversion rates.

One of our portfolio companies, Amperity, literally combines different silos of customer data. Companies that have customer data stored in disparate systems and tools can’t easily leverage this data to get a full picture of their customer base. Intelligently stitching these silos together drives significant business results for Amperity’s clients who can now clearly see the stitched 360-view of their customers and use it to market and sell products in a more intelligent way.

Trends Converge

Now is an exciting time for investors and entrepreneurs to be focusing on intelligent applications because of the momentum and growth of several important technology trends:

  • Massive computational power and low-cost storage are creating the infrastructure to train machine learning models
  • More data is generated and stored than ever before in many different fields like healthcare, autonomous systems, and media
  • Availability of good-enough capabilities at the edge to do a lot of the inferencing work at the edge as opposed to having to round-trip to the cloud
  • Continued improvement and development in tools and frameworks make it easier for companies and developers to begin using machine learning
  • New “user interfaces” using voice, vision, and touch are bridging the gap between the digital and physical world

As these trends make it easier for entrepreneurs to build intelligent applications, we have been developing our own frameworks to understand how all of these pieces fit together to create value for customers. Generally, we think about the intelligent application ecosystem in three main parts:

  • The Data Platform Layer
  • The Machine Learning Platform Layer
  • The Intelligent Applications and “Finished Services” Layer

The Machine Learning Platform Layer

As an early believer in the potential of AI and machine learning, Madrona has made several investments in the machine learning platform layer, including companies like Turi, Lattice, and Algorithmia. This layer of the intelligent app stack is meant to make it easier for other developers and applications to make use of machine learning by providing the tools and automating tasks such as model training, model deployment, and model management.

The ML platform includes machine learning frameworks like TensorFlow and PyTorch, managed services and tools like Amazon Sagemaker and TVM, as well as “Model as a Service” providers in the form something like AWS Marketplace that can help developers and companies develop and deploy ML models in specific environments. While many of these tools have been developed by large companies or acquired by large companies, we believe there continues to be interesting opportunities at this layer because deploying and managing machine learning systems continues to be very difficult.

As an example, while the major cloud providers have made large investments in software and hardware to train ML models in the cloud, using those models for inference at the edge continues to be a difficult problem on resource-constrained devices. Xnor.ai is a portfolio company in this segment that uses software optimizations to improve the quality of machine learning predictions on edge devices that have limited power or bandwidth.

Overall, we believe that while frameworks and tools have been improving, advanced techniques like reinforcement learning still need frameworks and tools that are easier to use, and there are many interesting opportunities to continue improving the ML platforms that intelligent apps depend on.

The Data Platform Layer

A precursor to using AI effectively and building intelligent applications is having a “data” strategy. Having a unique data strategy that could be a combination of public data sets and proprietary data sets enables companies to provide unique and differentiated value. This is a necessary first step, before you can use the data to train models and build a continuous learning system that is a core part of building an intelligent application.

Within the Data Platform layer, we think of companies and products from portfolio companies, Datacoral and Snowflake, as well as those from Databricks and Amazon’s Redshift, which offer customers different ways to connect, transform, warehouse, and analyze data in order to be used in an ML platform. What we’ve seen at this layer of the stack is that getting data into the right place, in the right format, in order to be used for machine learning continues to be very difficult, and simplifying this process is extremely valuable to customers.

Additionally, access and ownership to data itself is a key part of the data platform layer. By this, we mean that companies need to be thoughtful about their data strategies in order to find ways to gain access to, generate, or combine different data sources in order to create unique data assets. As we are seeing frequently in the news these days, companies also need to be thoughtful about data privacy and making sure customers understand what data is being used, shared, and how.

The lines between the Data Platform Layer, the ML Platform Layer, and Intelligent Apps themselves can be quite blurry, especially as companies try to offer their customers a broader set of services or learn their way into new customer needs. However, we do see a distinction between companies that are focused on helping customers manage their data vs. helping customers manage their ML models.

 

Ultimately, we are looking for companies that can benefit from the virtuous data cycle – where more data creates better user experiences, leading to better user engagement, leading to more data, and ultimately better user experiences again.

Intelligent Applications and “Finished Services” Layer

Within the Intelligent Applications and Finished Services layer, there are several ways to segment the market. We like to think about verticals – applications that focus on a specific industry such as healthcare or insurance – and horizontals – cross-industry applications such as marketing automation or robotic process automation. One of the principles that we follow when looking for these types of opportunities is to find areas where data is becoming digitized and/or more data is being collected than ever before.

For example, one promising vertical for intelligent apps is healthcare. Technology and regulatory trends have driven the healthcare field to rapidly digitize many different types of records – from basic medical histories, to insurance claims, to x-rays, MRI scans, and ‘omics’ data (e.g., genomics, proteomics, biomics). This digitization of healthcare is creating new levels of visibility into patient and population health data, and ML will be a critical tool to help decision makers make sense of these new data sources.

Workforce productivity is another promising area for horizontal intelligent applications because more data is digitized than ever before in HR and employee engagement across industries. One example of a horizontal intelligent app is Madrona Venture Labs spinout company, UpLevel, which uses unstructured data from tools like Slack to help managers get better insights on how to best engage their teams and drive productivity.

In addition to vertical and horizontal apps for business users, we also include other types of “finished services” in this bucket. This can include services like Amazon Rekognition or Amazon Forecast, which help application developers add image and video analysis or time series forecasting models to other applications. In this case, the end customer for a product may not be a consumer, but the product is a “finished service” which can be plugged into a customer-facing application.

In each of these use cases, we are looking to find companies that deeply understand customer pain points and use machine learning as a tool to solve customer problems, rather than starting with a technology and searching for use cases.

Areas of Opportunity

We believe that every successful application built today will be an intelligent application, and that is why we think there is a huge amount of opportunity for entrepreneurs in this space. In particular, we would love to see more companies that are building at the nexus of multiple large markets, companies with unique data strategies, and companies with great ML teams (because AI continues to be very difficult). Four specific areas where we are excited to meet new companies are:

  • AI for Healthcare – More healthcare data is digitized and stored than ever before, and this is creating massive opportunities to reduce costs while improving quality of care and operations. The intersection of the biological sciences with computer science is going to be a difficult area to break through, but the potential value created will be huge, and we are looking for entrepreneurs who are ready to take on these challenges.
  • AI for Work – More and more, companies want to measure and become data-driven about productivity, hiring, and employee wellness. Traditionally, HR and workforce data has been incredibly hard to collect and analyze, but new applications like Slack and Workday are creating opportunities for startups like Polly and UpLevel to analyze workplace data to generate insights for employees and managers.
  • Automation – Robotic Process Automation (RPA) vendors are one set of companies building early intelligent apps that can analyze a business process and improve productivity through automation, but they will not be the last. We think there will also be opportunities to build vertical “RPA-like” businesses in specific industries, automation of manual work that can be dangerous and expensive, and new types of autonomous systems like autonomous vehicles.
  • “End-to-End AI” – Many companies have a section of their pitch explaining how valuable their data will be. We always encourage companies to think about the best use cases for their data, and, if it makes sense, execute on those use cases themselves. Some of our favorite examples in this category are companies like Climate Corp, which started with an ML system for predicting weather, found that they could use their predictions to sell weather insurance to farms, and eventually built an end-to-end farm management software system to capture more data and use it to write insurance policies.

Conclusion

During a recent CIO roundtable, we debated whether machine learning was an over-hyped or under-hyped technology trend. The answer in most people’s minds was both. There are incredibly high expectations for machine learning, and many of those expectations are not grounded in the reality of what ML can do today.

However, we believe that as we move forward, the ability to build new applications and continuously improve systems and processes using machine learning will be a core part of any app, and machine learning will be immensely impactful in every fabric of the society that we work and live in.

Current or previous Madrona Venture Group portfolio companies mentioned in this blog post: Algorithmia, Amperity, Datacoral, Lattice, Snowflake, Suplari, Turi, Xnor.ai, UIpath

The Remaking of Enterprise Infrastructure – Investment Themes For Next Generation Cloud

Enterprise infrastructure has been one of the foundational investment themes here at Madrona since the inception of the firm. From the likes of Isilon to Qumulo, Igneous, Tier 3, and to Heptio, Snowflake and Datacoral more recently, we have been fortunate to partner with world-class founders who have reinvented and redefined enterprise infrastructure.

For the past several years, with enterprises rapidly adopting cloud and open source software, we have primarily focused on cloud-native technologies and developer-focused services that have enabled the move to cloud. We invested in categories like containerization, orchestration, and CI/CD that have now considerably matured. Looking ahead, with cloud adoption entering the middle innings but with technologies such as Machine Learning truly coming into play and cloud native innovation continuing at a dizzying pace, we believe that enterprise infrastructure is going to get reinvented yet again. Infrastructure, as we know it today, will look very different in the next decade. It will become much more application-centric, abstracted – maybe even fully automated – with specialized hardware often available to address the needs of next-generation applications.

As we wrote in our recent post describing Madrona’s overall investment themes for 2019, this continued evolution of next-generation cloud infrastructure remains the foundational layer of the innovation stack against which we primarily invest. In this piece, we go deeper into the categories that we see ourselves spending the most time, energy and dollars over the next several years. While these categories are arranged primarily from a technology trend standpoint (as illustrated in the graphic above), they also align with where we anticipate the greatest customer needs for cost, performance, agility, simplification, usability, and enterprise-ready features.

Management of cloud-native applications across hybrid infrastructure

2018 was undeniably the year of “hybrid cloud.” AWS announced Outposts, Google released GKE On-Prem and Microsoft beefed up Azure Stack (first announced in late 2017). The top cloud providers officially recognized that not every workload will move to the cloud and that the cloud will need to go to those workloads. However, while not all computing will move to public clouds, we firmly believe that all computing will eventually follow a cloud model, offering automation, portability and reliability at scale across public clouds, on-prem and every hybrid variation in between.

In this “hybrid cloud forever” world businesses want more than just the ability to move workloads between environments. They want consistent experiences so that they can develop their applications once and run anywhere with complete visibility, security and reliability — and have a single playbook for all environments.

This leads to opportunities in the following areas:

  • Monitoring and observability: As more and more cloud-native applications are deployed in hybrid environments, enterprises will demand complete monitoring and observability to know exactly how their applications are running. The key will be to offer a “single pane of glass” (complete with management) across multiple clouds and hybrid environments, thereby building a moat against the “consoles” offered by each public cloud provider. More importantly, the next-generation monitoring tools will need to be intelligent in applying Machine Learning to monitor and detect – potentially even remediate – error conditions for applications running across complex, distributed and diverse infrastructures.
  • SRE for the masses: According to Joe Beda, the co-founder of Heptio, “DevOps is a cultural shift whereby developers are aware of how their applications are run in a production environment and the operations folks are aware and empowered to know how the application works so that they can actively play a part in making the application more reliable.” The “operations” side of the equation is best exemplified by Google’s highly trained (and compensated) Site Reliability Engineers (SRE’s). As cloud adoption further matures, we believe that other enterprises will begin to embrace the SRE model but will be unable to attract or retain Google SRE level talent. Thus, there will be a need for tools that simplify and automate this role and help enterprise IT teams become Google-like operators with the performance, scalability and availability demanded by enterprise applications.
  • Security, compliance and policy management: Cloud, where enterprises lose total control over the underlying infrastructure, places unique security demands on cloud-native applications. Security ceases to be an afterthought – it now must be designed into applications from the beginning, and applications must be operated with the security posture front and center. This has created a new category of cloud native security companies that are continuing to grow. Current examples include portfolio company, Tigera, which has become the leader in network security for Kubernetes environments, and container security companies like Aqua, StackRox and Twistlock. In addition, data management and compliance – not just for data at rest but also for data in motion between distributed services and infrastructures – create a major pain point for CIOs and CSOs. Integris addresses the significant associated privacy considerations, partly fueled by GDPR and its clones. The holy grail is to analyze data without compromising privacy. Technologies such as security enclaves and blockchains are also enabling interesting opportunities in this space and we expect to see more.
  • Microservices management and service mesh: With applications increasingly becoming distributed, open source projects such as Istio (Google) and Envoy (Lyft) have emerged to help address the great need to efficiently connect and discover microservices. While Envoy has seen relatively wide adoption, it has acted predominantly as an enabler for other services and businesses such as monitoring and security. With next-generation applications expected to leverage the best-in-class services, regardless of which cloud/on-prem/hybrid infrastructure they are run on, we see an opportunity to provide a uniform way to connect, secure, manage and discover microservices (run in a hybrid environment).
  • Streams processing: Customers are awash in data and events from across these hybrid environments including data from server logs, network wire data, sensors and IoT devices. Modern applications need to be able to handle the breadth and volume of data efficiently while delivering new real time capabilities. The area of streams processing is one of the most important areas of the application stack enabling developers to unlock the value in these sources of data in real time. We see fragmentation in the market across various approaches (Flink, Spark, Storm, Heron, etc.) and an opportunity for convergence. We will continue to watch this area to understand whether a differentiated company could be created.

Abstraction and automation of infrastructure

While containerization and all of the other CNCF projects promised simplification of dev and ops, the reality has turned out to be quite different. In order to develop, deploy and manage a distributed application today, both dev and ops teams need to be experts in a myriad of different tools, all the way from version control, orchestration systems, CI/CD tools, databases, to monitoring, security, etc. The increasingly crowded CNCF roadmap is a good reflection of that growing complexity. CNCF’s flagship conference, Kubecon, was hosted in Seattle in December and illustrated both the interest in cloud native technologies (attendees grew 8x since 2016 to over 8,000) as well as the need for increased usability, scalability, and help moving from experimentation to production. As a result, in the next few years, we anticipate that an opposite trend will take effect. We expect infrastructure to become far more “abstracted,” allowing developers to focus on code and letting the “machine” take care of all the nitty gritty of running infrastructure at scale. Specifically, we think opportunities are becoming available in the following areas:

  • Serverless becomes mainstream: For way too long, applications (and thereby developers) have remained captive of the legacy infrastructure stack in which applications were designed to conform to the infrastructure and not the other way around. Serverless, first introduced by AWS Lambda, broke that mold. It allowed developers to run applications without having to worry about infrastructure and to combine their own code with best-in-class services from others. While this has created a different concern for enterprises – applications architected to use Lambda can be difficult to port elsewhere – the benefits of serverless, in particular rapid product experimentation and cost, will compel a significant portion of the cloud workloads to adopt it. We firmly believe that we are at the very beginning of serverless adoption and we expect to see a lot more opportunities in this space to further facilitate serverless apps across infrastructure, similar to Serverless.com (toolkit for building serverless apps on any platform) and IOpipe (monitoring for serverless apps).
  • Infrastructure backend as code: The complexity of building distributed applications often far exceeds the complexity of the app’s core design and wastes valuable development time and budget. For every app, a developer wants to build, s/he ends up writing the same low-level distributed systems code again and again. We believe that will change and that the distributed systems backend will be automatically created and optimized for each app. Companies like Pulumi and projects like Dark are already great examples of this need.
  • Fully autonomous infrastructure: Automating management of systems has been the holy grail since the advent of enterprise computing. However, with the availability of “infinite” compute (in the cloud), telemetry data, and mature ML/AI technology, we anticipate significant progress towards the vision of fully autonomous infrastructure. Even in the case of cloud services, many complex configuration and management choices need be made to optimize the performance and costs of several infrastructure categories. These choices range from capacity management in a broad range of workloads to more complex decisions in specific workloads such as databases. In databases, for example, there has been some very promising research done on applying machine learning to basic configuration all the way to index maintenance. We believe there are exciting capabilities to be built and potentially new companies to be grown in this area.

Specialized infrastructure

Finally, we believe that specialized infrastructure will make a comeback to keep up with the demands of next-general application workloads. We expect to see that in both hardware and software.

  • Specialized hardware: While ML workloads continue to proliferate and general-purpose CPUs (and even GPUs) struggle to keep up, new specialized hardware has arrived from Google’s TPUs to Amazon’s new Inferentia chips in the cloud. Microsoft Azure also now offers FPGA-based acceleration for ML workloads while AWS offers FPGA accelerators that other companies can build upon – a notable example being the FPGA-based genomics acceleration built by Edico Genome. While we are unlikely to invest in a pure hardware company, we do believe that the availability of specialized hardware in the cloud will enable a variety of new investable applications involving rich media, medical imaging, genomic information, etc. that were not possible until recently.
  • Hardware-optimized software: With ML coming to every edge device – sensors, cameras, cars, robots, etc. – we believe that there is an enormous opportunity to optimize and run models on hardware endpoints with constrained compute, power and/or bandwidth. Xnor.ai, for example, optimizes ML models to run on resource-constrained edge devices. More broadly, we envision opportunities for software-defined hardware and open source hardware designs (such as RISC-V) that enable hardware to be rapidly configured specifically for various applications.

Open Source Everywhere

For every trend in enterprise infrastructure, we believe that open source will continue to be the predominant delivery and license mechanism. The associated business model will most likely include a proprietary enterprise product built around an open core, or a hosted service where the provider runs the open source as a service and charges for usage.

Our own yardstick for investing in open source-based companies remains the same. We look for companies based around projects that can make a single developer look like a “hero” by making her/him successful at some important task. We expect the developer mindshare for a given open source project to be reflected in metrics such as Github stars, growth in monthly downloads, etc. A successful business then can be created around that open source project to provide the capabilities that a team of developers and eventually an enterprise would need and pay for.

Conclusion

These categories are the “blueprints” we have in our minds as we look for the next billion-dollar business in the enterprise infrastructure category. Those blueprints, however, are by no means exhaustive. The best founders always surprise us by their ability to look ahead and predict where the world is going, before anyone else does. So, while this post describes some of the infrastructure themes we are interested in at Madrona, we are not exclusively thesis-driven. We are primarily founder driven; but we also believe that having a thoughtful point of view about the trends driving the industry – while being humble, curious and open-minded about opportunities we have not thought as deeply about – will enable us to partner with and help the next generation of successful entrepreneurs. So, if you have further thoughts on these themes, or especially are thinking about building a new company in any of these areas, please reach out to us!

Current or previous Madrona Venture Group portfolio companies mentioned in this blog post: Datacoral, Heptio, Igneous, Integris, IOpipe, Isilon, Pulumi, Qumulo, Snowflake, Tier 3, Tigera and Xnor.ai

Investment Themes for 2019

2018 was a busy year for Madrona and our portfolio companies. We raised our latest $300 million Fund VII, and we made 45 investments totaling ~$130 million. We also had several successful up-rounds and company exits with a combined increase of over $800 million in fund value and over $600 million in investor realized returns. We don’t see 2019 letting up, despite the somewhat volatile public markets. Over the past year we have continued to develop our investment themes as the technology and business markets developed and we lay out our key themes here.

For the past several years, Madrona has primarily been investing against a 3-layer innovation stack that includes cloud-native infrastructure at the bottom, intelligent applications (powered by data and data science) in the middle, and multi-sense user interfaces between humans and content/computing at the top. As 2019 kicks off, we thought it would be helpful to outline our updated, 4-layer model and highlight some key questions we are asking within these categories to facilitate ongoing conversations with entrepreneurs and others in the innovation economy.

For reference, we published our investment themes in previous years and our thinking since then has both expanded and become more focused as the market has matured and innovation has continued. A quick scan of this prior post illustrates our on-going focus on cloud infrastructure, intelligent applications, ML, edge computing, and security, as well as how our thinking has evolved.

Opportunities abound within AND across these four layers. Infinitely scalable and flexible cloud infrastructure is essential to train data models and build intelligent applications. Intelligent applications including natural language processing models or image recognition models power the multi-sense user interfaces like voice activation and image search that we increasingly experience on smartphones and home devices (Amazon Echo Show, Google Home). Further, when those services are leveraged to help solve a physical world problem, we end up with compelling end-user services like Booster Fuels in the USA or Luckin Coffee in China.

The new layer that we are spending considerable time on is the intersection between digital and physical experiences (DiPhy for short), particularly as it relates to consumer experiences and health care. For consumers, DiPhy experiences address a consumer need and resolve an end-user problem better than a solely digital or solely physical experience could. Madrona companies like Indochino, Pro.com and Rover.com provide solutions in these areas. In a different way, DiPhy is strongly represented in Seattle at the intersection of machine learning and health care with the incredible research and innovations coming out of the University of Washington Institute for Protein Design, the Allen Institute and the Fred Hutch Cancer Research Center. We are exploring the ways that Madrona can bring our “full stack” expertise to these health care related areas as well.

While continuing to push our curiosity and learning around these themes, they are guides not guardrails. We are finding some of the most compelling ideas and company founders where these layers intersect. Current company examples include voice and ML applied to the problem of physician documentation into electronic medical records (Saykara), integrating customer data across disparate infrastructure to build intelligent customer profiles and applications (Amperity), or cutting edge AI able to run efficiently in resource constrained edge devices (Xnor.ai).

Madrona remains deeply committed to backing the best entrepreneurs, in the Pacific NW, who are tackling the biggest markets in the world with differentiated technology and business models. Frequently, we find these opportunities adjacent to our specific themes where customer-obsessed founders have a fresh way to solve a pressing problem. This is why we are always excited to meet great founding teams looking to build bold companies.

Here are more thoughts and questions on our 4 core focus areas and where we feel the greatest opportunities currently lie. In subsequent posts, we will drill down in more detail into each thematic area.

Cloud Native Infrastructure

For the past several years, the primary theme we have been investing against in infrastructure is the developer and the enterprise move to the cloud, and specifically the adoption of cloud native technologies. We think about “cloud native” as being composed of several interrelated technologies and business practices: containerization, automation and orchestration, microservices, serverless or event-driven computing, and devops. We feel we are still in the early-middle innings of enterprise adoption of cloud computing broadly, but we are in the very early innings of the adoption of cloud native.

2018 was arguably the “year of Kubernetes” based on enterprise adoption, overall buzz and even the acquisition of Heptio by VMware. We continue to feel cloud native services, such as those represented by the CNCF Trail Map, will produce new companies supporting the enterprise shift to cloud native. Other areas of interest (that we will detail in a subsequent post) include technologies/services to support hybrid enterprise environments, infrastructure backend as code, serverless adoption enablers, SRE tools for devops, open source models for the enterprise, autonomous cloud systems, specialized infrastructure for machine learning, and security. Questions we are asking here include how the relationship between the open source community and the large cloud service providers will evolve going forward and how a broad-based embrace of “hybrid computing” will impact enterprise customer product/service needs, sales channels and post-sales services.

For a deeper dive click here.

Intelligent Applications with ML & AI

The utilization of data and machine learning in production has probably been the single biggest theme we have invested against over the past five years. We have moved from “big data” to machine learning platform technologies such as Turi, Algorithmia and Lattice Data to intelligent applications such as Amperity, Suplari and AnswerIQ. In the years ahead, “every application is intelligent” will likely be the single biggest investment theme, as machine learning continues to be applied to new and existing data sets, business processes, and vertical markets. We also expect to find interesting opportunities in services that enable edge devices to operate with intelligence, industry-specific applications where large amounts of data are being created like life sciences, services to make ML more accessible to the average customer, as well as emerging machine learning methodologies such as transfer learning and explainable AI. Key questions here include (a) how data rights and strategies will evolve as the power of data models becomes more apparent and (b) how to automate intelligent applications to be fully managed, closed loop systems that continually improve their recommendations and inferences.

For a deeper dive click here.

Next Generation User Interfaces

Just as the mouse and touch screen ushered in new applications for computing and mobility, new modes of computer interaction like voice and gestures are catalyzing compelling new applications for consumers and businesses. The advent of Alexa Echo and Show, Google Home, and a more intelligent Siri service have dramatically changed how we interact with technology in our personal lives. Limited now to short simple actions, voice is becoming a common approach for classic use cases like search, music discovery, food/ride ordering and other activities. Madrona’s investment in Pulse Labs gives us unique visibility into next generation voice applications in areas like home control, ecommerce and ‘smart kitchen’ services. We are also enthused about new mobile voice/AR business applications for field service technicians, assisted retail shopping (E.g., Ikea’s ARKit furniture app) and many others including medical imaging/training.

Vision and image recognition are also rapidly becoming ways for people and machines to interact with one another as facial recognition security on iPhones or intelligent image recognition systems highlight. Augmented and virtual reality are growing much more slowly than initially expected, but mobile phone-enabled AR will become an increasingly important tool for immersive experiences, particularly visually-focused vocations such as architecture, marketing, and real estate. “Mobile-first” has become table stakes for new applications, but we expect to see more “do less, but much better” opportunities both in consumer and enterprise with elegantly designed UIs. Questions central to this theme include (a) what ‘high-value’ new experiences are truly best or only possible when voice, gesture and the overlay of AR/VR/MR are leveraged? (b) what will be the limits of image (especially facial recognition) in certain application areas, (c) how effective can image-driven systems like digital pathology be at augmenting human expertise, and (d) how will multi-sense point solutions in the home, car and store evolve into platforms?

For a deeper dive click here.

DiPhy (digital-physical converged customer experiences)

The first twenty years of the internet age were principally focused on moving experiences from the physical world to the digital world. Amazon enabled us to find, discover and buy just about anything from our laptops or mobile devices in the comfort of our home. The next twenty years will be principally focused on leveraging the technologies the internet age has produced to improve our experiences in the physical world. Just as the shift from physical to digital has massively impacted our daily lives (mostly for the better), the application of technology to improve the physical will have a similar if not greater impact.

We have seen examples of this trend through consumer applications like Uber and Lyft as well as digital marketplaces that connect dog owners to people who will take care of their dogs (Rover). Mobile devices (principally smartphones today) are the connection point between these two worlds and as voice and vision capabilities become more powerful so will the apps that reduce friction in our lives. As we look at other DiPhy sectors and opportunities, one where the landscape will change drastically over the coming decades is physical retail. Specifically, we are excited about digital native retailers and brands adding compelling physical experiences, increasing digitization of legacy retail space, and improving supply chain and logistics down to where the consumer receives their goods/services. Important questions here include (a) how traditional retailers and consumer services will evolve to embrace these opportunities and (b) how the deployment of edge AI will reduce friction and accelerate the adoption of new experiences.

For a deeper dive click here.

We look forward to hearing from many of you who are working on companies in these areas and, most importantly, to continuing the conversation with all of you in the community and pushing each other’s thinking around these trends. To that end, over the coming weeks we will post a series of additional blogs that go into more depth in each of our four thematic areas.

Matt, Tim, Soma, Len, Scott, Hope, Paul, Tom, Sudip, Maria, Dan, Chris and Elisa

(to get in touch just go to the team page – our contact info is in our profiles)

The State of Today’s Autonomous Vehicle Market


Only five years ago, the autonomous vehicle future seemed like a distant vision.
This year at CES, the “frenzy” over autonomous vehicles stole the show with dozens of live demos, partnerships, and product announcements. In fact, 13 of the world’s 14 largest automakers have announced plans to bring autonomous vehicles to market, and 12 of the world’s 14 largest technology companies have announced plans to build technologies to support and operate autonomous vehicles.

This sudden interest in autonomous driving has bid up the “going rate” for autonomous driving talent to, at least in one case, $10 million per person (Harvard, MIT, CMU, and Stanford students take note!)

Clearly, many large companies are investing heavily in autonomous driving technology because they see that autonomous cars not only have the ability to drastically change the auto industry but also see the enormous cultural change that the presence of AVs could create. If we spend more time in the car and have more time there to do stuff . . they want to to be there.

As investors we have been most interested in watching how different stakeholders are shaping their strategies for competing in this market. For example, will automakers build their own autonomous technology, rely on partners, or both? Do automakers think they will continue selling autonomous vehicles to consumers or only to ride-sharing services? Will ride-sharing services companies want to move into other areas of the industry including building autonomous technologies?

The Ever Changing Autonomous Vehicle Value Chain

Here is our view of the current state of the autonomous vehicle value chain. The three high-level layers of the autonomous vehicle market are

  • Service providers (e.g., ride-hailing, ride-sharing, rentals),
  • Technology providers (both hardware and software), and
  • Automobile manufacturing


We see a significant number of companies sitting in between layers – such as Tesla building an autonomous driving system as well as being a car manufacturer – and we also see companies that have historically operated in one area making large investments of capital or time in other layers in order to move into other areas of the value chain. An example of this is

ReachNow, BMW’s rental and ride hailing service which is, among other things, a hedge against being cut out of a possible future where ride hailing vs car owning is the norm. Each of these companies sees an opportunity to capture a larger portion of the end-state autonomous vehicle value chain for themselves, and they are positioning themselves accordingly.

The AV Partnership Matrix – De-risking the future by teaming up

In addition to investing or acquiring companies or talent, companies have also begun forming partnerships to ensure they do not get cut out of valuable portions of the autonomous vehicle market or caught with single source suppliers for key technologies.

For example, with the rise of autonomous vehicles, companies realized that detailed maps might (though it is still under debate) be one of the most critical inputs to self-driving cars for determining whether the car is seeing the environment or another vehicle, person, or object in its environment. This led to several big moves in developing in-house maps and/or acquiring access to other sets of mapping data. For example, in August of 2015, a consortium of automakers bought Here maps for $3 billion; in July of 2016, Uber announced a $500 million plan to map the world’s roads; and in December of 2016, Mobileye announced a partnership with Here’s owners to share their mapping data.

Some other interesting takeaways we see from looking across the autonomous vehicle landscape are:

  • The companies working with the major technology providers are also developing their own homegrown systems. For example, while Volvo is providing Uber with vehicles for their well-publicized Pittsburgh, San Francisco, and Arizona tests, they are developing their own autonomous systems as well through their Drive Me research project.
  • There have already been some high-profile ‘breakups’ in the autonomous vehicle space. After Tesla’s crash, Tesla and Mobileye have pointed several fingers on who fired who (and whose technology led to a fatal crash), Baidu and BMW called off their joint work citing different development paces and ideas about research; but over time we will likely see more differences in tech and/or business philosophies that will lead companies to go their separate ways.
  • Most of the major automakers have aligned themselves with a ride-hailing service – either via investment in the case of Toyota-Uber or GM-Lyft or in other cases building it in house or making a full acquisition. These moves have been interesting because the strength of Uber or Lyft’s driver network becomes less relevant in an autonomous, ride-hailing future. When you can put cars on the road without a person at the wheel, different elements become more important levers of success – namely the ability to manufacture, finance, and maintain cars. This could give automakers a head start later in the game, so to speak. Though success in ride hailing is also dependent on consumer penetration so companies like Uber and Lyft might have their own leverage over the automaker latecomers.

Other Stakeholders – What are the regulators, drivers, and consumers doing?

As we continue watching the moves that software companies and automotive companies are making in the autonomous vehicle space, we have also been keeping track of what regulators, drivers, and consumers are thinking and saying about autonomous vehicles.

To date, we have been impressed with how proactive the federal, state, and local governments have been in their support of autonomous driving technology. It appears regulators and planners at multiple levels are bought into the potential promise of fewer accidents, less congestion, more productive time for citizens and freed up space now devoted to parking lots in cities. But they are treading carefully, encouraging companies to experiment in safe, controlled ways. When the US NHTSA investigated the fatal crash involving Tesla’s Autopilot – the findings were actually that, though it failed in that instance, Autopilot had decreased the amount of crashes by 40% since introduction of the technology.

However, as this technology becomes mainstream and moves from high-end cars to widespread adoption among ride-hailing and trucking companies, there could be massive disruption to the way the workforce is structured in many different places around the country. A 2015 NPR review of Census data shows that Truck Driver is the most common job in nearly every US state. Autonomous trucks will have a massive impact on the trucking industry.

As more startups and large companies begin public demonstrations and public releases of their products, they must find the right ways to introduce these technologies for both public safety and public perception. “Drive fast and break things” will not be the right approach to releasing autonomous vehicles, and companies need to be thoughtful about the best way to introduce these technologies.

As investors (and eager consumers and citizens) we are watching how the AV market is evolving and looking for opportunities. If the innovations in the last five years happened twice as fast as expected, imagine where we could be in another five – or maybe just two and a half!