April 25, 2017 –
The 7 Major Funding Pitfalls and Windfalls for 2017
2016 marked the passings of some of my beloved celebrities (George Michael, immediately followed by Princess Leia?!), as well as the Presidential election turmoil, and many of us were happy to see the year end— but we’re now faced with even more uncertainty in 2017, and wondering what lies ahead. At Mayfield, we started the year wrestling with what the new presidential administration means for innovators, for entrepreneurs, for investors, and for the market. After taking stock of 2016, and breathing deeply through the first quarter of 2017, I’ve looked ahead and here are some things to keep in mind:
#1 Remember That Good Years Can Bring in the Subsequent Bad Years (and We’re Seemingly Overdue to the Next Downswing!)
There’s strong evidence that more money is still flowing into the VC and startup ecosystem — and that it’s coming from fewer sources. We see rising commitments coincident with significant consolidation. The question, then, is whether more money is a good thing. It’s arguable that an easy market foretells a tough one. It’s hard to turn down billions of dollars from limited partners, but the most disciplined of investment funds know that the work we do in venture capital doesn’t, itself, really scale. We couldn’t deploy that much more effectively, especially at the early stage, but it’s also very tempting when limited partners want to invest more into venture funds, reading about unicorn valuations and blockbuster deals they missed out on previously. It’s also healthy to remind ourselves that taking in too much money likely means taking more bets, with larger checks at higher valuations, and those bets may very well lead to destabilizing the market by raising the noise level and subsequently lowering overall industry returns.
#2 Undercapitalization Will Be Your Archenemy
It’s easy to raise a seed round right now, as Seed is the New Series A (and “pre-seed” is the new Seed round!). The danger is undercapitalizing an early stage company, and the resulting limbo of “tweener traction” is a startup founder’s main enemy. If you only raise $750k, then you buy yourself at most 12 to 16 months. The challenge then becomes the likely ensuing Series A crunch: It’s increasingly common for VCs to say, “Hmm, the 100K users you’ve reached with your seed round is promising … but come back when you have more data and are closer to 1M users.” That leads to “bridge” seeds: e.g., Seed-2 extensions of the original Seed-1 round. A Seed-2 is really a bridge round, in essence a failed Series A. Seed funds aren’t historically setup to bridge their existing seed deals, so founders have to often open up their prior round (or worse yet, take a down-round vs. the prior convertible note cap, and/or agree to a lower priced round). The lesson is you need to be capitalized sufficiently early on so that you don’t waste precious cycles fundraising, and so you don’t risk telegraphing a poor impression to the market. I expect to see seed rounds aiming out for longer 18-24 month runways before needing to raise A rounds.
#3 AR/VR Is Hot, and the Odd Bets Will Be the Best Bets
AR/VR is all the rage (hype?) right now. When you study the source of funding, the seeds and the Series As are from classic funds … but the Series B and Series C are, by contrast, distinctly strategic and often from international players and platforms — and often at illogically high valuations. It’s almost never based on multiples of revenue because, overall, there’s not much real revenue in AR/VR yet. As an investor, I’m always considering what I think of as “time to revenue,” but in AR/VR land that time period is still especially uncertain. I think there are three cuts to consider:
- Cut 1: AR versus VR. AR is more enterprise, and shows more revenue.
- Cut 2: Mobile VR versus head-mounted VR. The difference is that with mobile VR you can you just use your phone, meaning there’s no install-base problem.
- Cut 3: Entertainment versus non-obvious and emergent applications. Entertainment is gaming, cinematic, social, etc. For new platforms, porting an older business model simply isn’t as good as an entirely new business model or use case that is unexpected but completely native to the new platform. Think about job training, architecture, data visualization, health care applications, and creative processes that are fundamentally transformed through AR and VR — that’s where the next big businesses will appear.
#4 AI and Machine Learning Startups Will Live/Die on Recruitment
AI and machine learning are also currently blessed and cursed with tons of interest and hype. These terms are now requisites tacked onto every pitch, yet can mean many different things in practice, from horizontal platforms (think Google) to specific verticals, like customer service, or sales, or personal consumer-side apps. These days, if you don’t have the words “AI” or “machine learning” in your deck, you may be perceived as behind the times. Hot business buzzwords are nothing new (e.g., “cloud,” “big data,” “mobile-first,” “social”). Now it’s assumed that every new startup must have “AI” and “machine learning” as a core competency, but my litmus test as an investor is as follows: how many PhDs from the six or so world-class labs are on your team? They’re the new rock stars, and these days can command a million dollar salary. If every company needs that talent, but the very best teams are being snapped up by the Facebooks, Amazons, and Googles of the world, the question becomes: where do we hire them from? Understand that a $300k starting salary (which is way beyond what most startups can afford to offer) is not enough, and these roles are even harder to hire for than developers and designers. And now that every business plan template is essentially “take existing application X and add AI/ML to enhance it,” we’re led back to same question: where will the necessary talent for these startups come from?
#5 Don’t Think AI; Think Workflow Automation
We all say AI/machine, but in these early days it seems to be more about workflow automation and enhancement via insights rather than general artificial intelligence platforms (which result in the classic “hammer seeking a nail” product-market fit problem). Early revenues for AI startups are most often found in vertical and departmental enterprise applications, in which you have corporations with massive proprietary datasets, existing apps and workflows, and the need to garner more insights and efficiency with these. For example, our portfolio company Outreach.io offers a plugin on top of existing emails and CRM. The system analyzes every sales email and can even offer to craft a more effective message for salespeople before they hit the send button, based on all the wisdom it has gained by studying thousands and thousands of sales email transactions.
There are three perceived steps to the AI-enabled economy to be kept in mind:
- Step 1: Machine assists human.
- Step 2: Humans assist machine.
- Step 3: Machines take human’s job.
That progression highlights the existential question of our entire generation. The last election in the U.S. was ultimately about the advancement of technology versus humans. Even if we keep the jobs here, they’re going to be automated at some point. If regulation is implemented badly, it could create artificially high levels of paid labor. Tech is seen as the perceived enemy, and maybe we are the enemy. I now think more than ever about these questions when making new investments: am I helping creating more jobs and new forms of “work,” like we have through Lyft? Will this technology and company improve humanity and life for people even outside of SF and NY? (Frankly, I’m also thinking that in 5 to 10 years, armed with enough training data, an AI will probably do my own job better than I can.)
#6 Don’t Reinvent the Car; Reinvent the Carburetor
As the drive to autonomous vehicles proceeds, it appears that smaller pieces of the car have been more successful than the cars themselves. At Mayfield we have some exposure here. We were the first VC in Lyft. What we’re observing is massive pre-IPO sales of smaller companies: Cruise to GM, Otto to Uber. These days if you have a whip-smart AI team from Stanford, they make a reference platform, and then they’re snapped up before they even have a product. The autonomous revolution will hit the market first in trucking. It’ll augment and then replace drivers. We’re looking at the pieces of the puzzles, and the car is one of the puzzles — like Delphi, a sensor package that is something you put on existing cars, which then benefit from AI software. I think that is an example where there is a lot of investing to be done.
#7 Digital Health Is Going to Be an Enterprise Play for the Foreseeable Future
It can help in all this prognostication to be clear about where I’ve been wrong in the past. I was wrong about digital health. I went deep into the consumer side, and it became clear that wasn’t a path to success. The good news is that the enterprise and payer side have woken up, and they have picked up the slack. Let’s talk, first, about where I was wrong. My theory was that, on the consumer side, companies would disrupt health care by engaging end users, by bypassing the medical establishment — by tracking behavior, changing what you eat, and so forth. HealthTap and Lantern are examples of companies that built great consumer-grade products, but the conversion turned out to be very hard due to price. This, however, got the attention of self-insured employers, and they started landing enterprise deals to give these benefits to employees. This then got insurance companies interested. Which is to say, it all worked out, but I started out with a mistaken impression. Well, maybe not entirely mistaken: The irony is that if companies like HealthTap and Lantern had started out with enterprise, they would already have probably run out of money. It ended up being a reverse hack: the disruptors are pivoting to get money from employers, and then payers and healthcare institutions — who are driven primarily by their innovation in applying consumer-oriented best practices, delight, beauty, and ease of use into their products! Digital Health startups that aren’t able to manage this dance of consumer, employer, and payer are likely to run out of steam in 2017, as they’ll be unable to demonstrate scalable business models.
There’s Always a Bubble (and a Pot of Gold) Somewhere
Despite some of my cautionary tone, I remain optimistic overall about 2017. We should see some pent-up IPOs in the backlog hit the market (hopefully to strong reception from the Street), and the existential threats of ecommerce, mobile, messaging-native usage, AI/ML, and automation should continue to push large incumbents to make significant strategic investments to keep pace with Facebook, Amazon, Google, Alibaba, Tesla, Netflix, Apple, and the like. And even if the market heavily corrects, this will present an opportunity for investors to jump in at highly reduced valuations — I made many of my best investments in the wake of the Global Financial Crisis, as those startups rode the recovery cycle in the subsequent years.
I’m often asked if the market is in a bubble. I tend to see the landscape as a poorly microwaved meal: the overall average temperature may be hotter or colder (and we’ve certainly seen an unusually long bull-market run now), but there are usually overheated spots here and there that are bubbling and popping, accompanied by several cold or even frozen pockets, which often represent contrarian places to play. (Consider semiconductors, which has long been a cold investment sector. It is ripe for uptick fueled by the need for specialty silicon for AI, machine vision, VR, autonomous vehicles, and other new applications.)
In the meantime, let’s hope that the turmoil in Washington, D.C., doesn’t introduce too much non-market systemic volatility into the markets — and that the Grim Reaper will lay off beloved celebrities this year!