Momentum and the Probability of Success
“The general who wins the battle makes many calculations in his temple before the battle is fought. The general who loses makes but few calculations beforehand.”
- Sun Tzu, The Art of War
When I set out to write about decision trees and power laws, my goal was to better understand the attributes in which great investors look for in early stage companies. The conclusion I came to was that all early stage companies have a high probability of failure, but the magnitude of the potential outcomes are vastly different. Vinod Khosla said it best when he said,
“Most technology startups fail. There’s a winner, and there’s 7 out of 10 that lose. I don’t mind failing, but if I succeed it better be worth succeeding for.”
The best early stage VC investments aren’t those with the highest probability of success, but rather a weighted average of probabilities (from a variety of characteristics) with a potential outcome that is so big it can return their funds. This was the basis behind the decision tree I laid out and was reinforced in my mind after listening to Chris Dixon on this podcast. During the interview Chris says,
“When you meet entrepreneurs you’re sort of thinking somewhat will they succeed, but probably thinking more about if they succeed, how big can it get?”
A day earlier I engaged with Mark Suster on Twitter who had sent out a tweetstorm about his skepticism when everybody is looking to fund a category all at once. It’s a fine line because historically some of the greatest technology companies were built directly after a new platform emerges, which excites investors and the ecosystem alike. It also opens up a host of white space and new opportunities that weren’t previously visible.
That can also be too obvious, whereas the number of people chasing the same opportunities creates what is taught in business schools as a “perfect market”, one in which drives returns towards the cost of capital. In early stage technology however, this means rising customer acquisition costs, margin pressure, and instead of driving returns towards the cost of capital you may wind up with one big winner and a whole host of failed companies.
For the vast majority of investors, predicting this winner is like playing the lottery. Some might argue for making the investment anyways because of reasons like first mover advantage or the fact that they’re building a portfolio of numerous companies. I personally don’t have enough experience to say whether or not that’s a worthwhile strategy.
I wrote earlier about the “expected value” formula in statistics (not a great term for venture since it’s hard to expect value) which is the “weighted-average value for a distribution of possible outcomes”. To use an exaggerated example, the market for time travel might be infiniteness, but the probability for a successful outcome at this time is zero.
A better strategy might be to look for a market where there is technological momentum as well as structural changes that creates opportunities and white space. It’s not a flawless strategy as the herd is always chasing these markets at the first sign of opportunity, which I previously called the Efficient Startup Hypothesis. To quote Sequoia founder Don Valentine,
“The key to making great investments is to assume that the past is wrong, and to do something that’s not part of the past, to do something entirely differently.”
Instead of speaking broadly, below I outlined a few markets in which I believe there is significant momentum, which may render past market dynamics obsolete for a variety of reasons:
- Healthcare — I’m not the first person to point this out, so it’s a bit hypocritical given my points above but the the opportunity here is massive and there’s real fundamental technological shift in addition to changing market dynamics. In the past providers were incentivized to provide “services” to patients and billed them accordingly. As the Bessemer Ventures team pointed out recently, this is rapidly changing and 75% or more of healthcare payments is expected to be “value-based” by 2020. Healthcare costs have also skyrocketed to astronomical proportions in the United States and there is a real shortage of primary care physicians. At the same time, we’re all carrying around a supercomputer in our pockets and it seems like every week a new digital device, app, or wearable is coming to market that monitors another aspect of our health. There is no reason that in the future we should go to the doctor only when we feel something is wrong, but rather remote monitoring, maintenance, and diagnosis should happen automatically and in the background without conscious thought. There are many other opportunities that are emerging in this space and given that it’s over 17% of our country’s GDP, I’d say it’s a pretty massive market to go after.
- Insurance — Similarly to healthcare, legacy underwriting models were designed for a different time where across all insurance products, risks were statistically modeled based on a variety of factors. Given the whole supercomputer in our pockets point mentioned above, data collected from online and offline interactions provides a massive opportunity for new underwriting, pricing, and products tailored towards are lives today vs 50–100 years ago. It won’t happen overnight as traditional insurance companies aren’t yet equipped to handle all of this data and it doesn’t make sense for them to completely change their models either. While today many startups are attempting to shift the paradigm and start their own insurance companies (through a Managing General Agent model), eventually legacy insurance companies will see these threats and realize that it behooves them to build out the necessary infrastructure that allows them to buy, integrate and build the right technology tools and services to continue their decades-long dominance. At the same time, access to new data points and new behaviors offers the opportunity for new insurance products vs past all-encompassing products.
- “The Rent is Too Damn High” — Given a shift in demographics, the rising population, and costs in cities, rent is increasing for all businesses in major markets. At the same time, consumer expectations and patience have changed dramatically and traditional brick and mortar stores are struggling due to Amazon and other online merchants (Macy’s recently announced the close of 14 stores after holidays offline sales figures came in). Like we’ve seen with the rise of direct to consumer commerce, owning your own distribution is not only possible due to companies like Google and Facebook providing access through ads to all smartphones users at all times, but it’s actually more profitable to do so and offers less focus on margins and more focus on product and the customer. Early successes leveraging this model like Munchery, Kitchensurfing, BlueApron/Plated, and even Uber to some extent, are just the beginning. There’s an opportunity for chefs, retailers, and others that provide a good (providing a service is harder due to unit economics as we’ve seen with the on-demand economy) to bypass the “wholesale transfer pricing problem” and sell their product directly to customers, leveraging the scale provided by the Internet, and better unit economics that lead to costs savings for the end consumer. One of my favorite short reads this year was written by Om Malik where he talked about replacing McDonalds with these models. It seems like every year customer acquisition via prime real estate and physical sprawl matters less and less and while the stores/restaurants that offer the best experiences aren’t going anywhere anytime soon, there’s a long tail that could provide much better value without the burden.
- Unbundling the employee from the organization — In the past, companies needed employees under the same roof to share costs, communicate in a timely and efficient manner, and provide access to secure company data. This is no longer the case and with intelligent search engines and access to any and all data with a quick query, we’re not far away from a world where it’s more efficient for companies to hire employees for specific tasks on a part time basis. The cultural impact will remain to be seen, but so many salary employees spend their time inefficiently because they work for one company vs doing one task that they do very well for many companies.
- Brick and Mortar Payments — Us millennials mostly pay each other using Venmo but the in-store experience is no where near as easy and simple. Transaction fees are too high for most small merchants and credit cards just aren’t secure. Apple, Google, Samsung and others are still trying to sure up their mobile payments strategy, but that doesn’t really solve the crux of these issues. Many of us thought bitcoin would be the solution but truthfully, there just hasn’t been enough market momentum to date to move the needle. However, this might be one of those “overestimate technology in the short term and underestimate it over the long term” moments. On one end of the spectrum, the global financial markets are in disarray. There’s a lot of fear and uncertainty which certainly makes a case for using bitcoin as a store of value. While the US banking system and the dollar don’t keep most Americans up at night, there’s another problem emerging that may lead to the adoption of bitcoin and the blockchain here in the US. As technology has advanced, more and more opportunities and tools for hackers and data breaches have become omnipresent. As a result, retailers lost $30B in chargebacks last year and paid $61B in interchange fees (which banks say helps them to cover the costs of fraud). Issuers are forcing retailers to upgrade their terminals and use chip readers instead of card swipes, a practice that is meant to reduce fraud but creates more friction for retailers as it slows up checkout times (pain for the consumers). I don’t know what the answer is and the space is certainly littered with dead bodies, but I do know that traditionally it takes up to 3 days to clear a transaction and just minutes on the blockchain. Will retailers begin to incentivize cryptocurrency payments? Maybe It’ll be bitcoin and the blockchain that solves this problem (or maybe it won’t) but there’s finally real market momentum and pain for numerous stakeholders, so it’s a market worth keeping an eye on.
I should point out that a lot of investors are thinking about these categories/trends so the fallacy that Mark Suster tweeted about isn’t unavoidable here. However, I believe the various markets discussed above have momentum and opportunity which increases the probability of success (even if by a minuscule amount) while still having massive potential payoff if success is achieved.
Understanding the intricacies of these markets and the fundamental market or technological shifts that offers an opportunity for a new startup to emerge is incredibly important. This is the fallacy that many incubators and accelerators fall victim to.
The game is predicated on finding ways to increase the probability of success while maintaining the potential size of that lower probability successful outcome. This is very different from trying on an Oculus Rift and deciding you want to invest in VR companies because it’s “the next big thing”. In this scenario, while the magnitude of the potential outcome might be high, you’re just rolling the dice along with so many others, but not really increasing your probability of success. Context matters.
I’m always looking to add and remove markets from this list so please challenge me and let me know what other markets you think fit the bill (@josh_nussbaum on Twitter).
Note: I didn’t talk about team here since it’s been written about at length, but it’s the number one way to increase a startup’s probability for success (although with the wrong market dynamics even the best teams may not find success).