There are many ways to estimate the expected value of a startup investment.
The quant approach. The expected value is the sum of: the probabilities of each possible outcome multiplied by the profit in that outcome. For example, see Bessemer’s venture memos - here’s one for Shopify.
But how do we estimate the probabilities? This is the core of angel investing skill. Although some statistics might help set initial upper and lower bounds on outcome distributions, experience and instinct play the biggest role here.
Profit is a function of current valuation and exit valuation, but how do we estimate exit valuation? Profit is also a function of how much you get diluted by future financing rounds, which depends in part on capital efficiency (more on this below).
In short, this is hard. My thesis is that two factors dominate the result: evidence of slight increases in the probability of not failing, and evidence of significant increases in the best case exit valuation.
Anchor to comparable public company exit values (or acquisitions) in the same sector. I shy away from this method though because the startup I’m evaluating today is aiming for an exit 5-10 years in the future when the whole market landscape may be different. Also, many of the biggest exits in tech have been for companies that had no comp early on. As Peter Thiel said, competition is for losers.
One special case - if a startup might be able to dominate a market from network effects, IP, or other barrier to competition, then I focus on whether the startup is doing something that will accelerate the total addressable market beyond its current natural growth rate (see, e.g., Bill Gurley’s article about Uber’s TAM). Note - this is not the same as growing the SAM (serviceable addressable market) by expanding into new customer bases, revenue streams, etc. As Aaron Levie put it:
Sizing the market for a disruptor based on an incumbent's market is like sizing the car industry off how many horses there were in 1910.In these cases, valuations and round sizes can start to look very large relative to the current revenue and TAM, but there is a method to the madness.
Here are some heuristics that I’ve developed around the inputs to #1 above:
LTV/CAC ratios (lifetime customer value / customer acquisition cost) prior to Series A are often worse than useless, except perhaps as a filter for companies where the ratio is less than 1.
The greatest companies often have organic growth early with almost no marketing spend (which sends this ratio absurdly high).
At the seed stage, LTV is a future estimate with massive error bars, while CAC is a current known quantity. But nobody ever presents the ratio as a range.
LTV and CAC can change significantly over time. Some businesses have a CAC that grows over time (e.g. first movers, as well-funded competitors enter, or a disruptive new consumer product/service that is more expensive to sell beyond early adopters), others will have a CAC that shrinks over time from viral effects, marketplace dynamics, or more efficient marketing from data collection/targeting. Some businesses can add new services over time for the same customers, or improve product quality to reduce churn, both of which increase LTV, while others will grow revenue faster by eating up a market with massive TAM (i.e. growing customers), but LTV will remain relatively steady, or even decline for pricier products/services if forced to move downmarket in order to maintain customer and revenue growth.
I want to back startups that are most likely to attract top tier VCs at Series A, which increases the odds of attracting top employees, other top VCs for future rounds, etc. Therefore, I need to balance that goal against my views on long term potential - they’re correlated, but may not be the same because some sectors go in and out of favor with VCs, and the company may not be able to generate enough KPI momentum to attract the best VCs (e.g. because they don’t have enough seed capital or headcount to scale fast).
Entry valuation at the seed stage doesn’t matter much because of the power law distribution of exit values. In public equity terms, startup investing is almost 100% momentum investing, not value investing. I want to find a few great companies at fair prices, not a bunch of fair companies at great prices. (One caveat - I’m assuming a lead investor has validated the valuation and that I’m not the first investor in the conversation.) Realizing a 10X 21return instead of a 5X return because you got in at half price is not going to make up for missing the 100X return company. The startup market is super illiquid, and at an early stage the startup supply is generally not fungible. Investors are rarely in a position to decline an investment in one company because they found another at the same time that is extremely similar at a lower price. Therefore, if a company receives two or three term sheets from potential leads instead of one, the valuation could end up massively higher (in some cases 50-100% higher) than the founder’s initial target. The existence of that auction dynamic is a signal that increases the probability of non-failure significantly.
Companies often use a multiple of annualized revenue as the basis for a valuation. I’ve learned to be careful with this even though it’s tempting to oversimplify. Some businesses are seasonal or subject to volatile revenue recognition, so trailing 12 months actual revenue and current annualized revenue can diverge a lot. Month over month sustainable growth rate is far more important (and at the Series A, will be a big factor in the revenue multiple for valuation).
There may be some general patterns in market revenue multiples based on sector - perhaps consumer goods are 3-5X, consumer subscriptions are 8-10X, B2B SaaS subscriptions with low churn are 12-15X, etc. But a consumer subscription business could get a 20X multiple if it’s growth rate is 30-40% month over month for a year instead of 10-20%.
By “sustainable”, I mean growth isn’t being driven by pricing to customers at negative gross margins, or through paid marketing that is far more expensive than the customer value, and that churn is within industry norms.
In a prior newsletter, I talked about founder perseverance. This can have quite a mighty impact on exit valuation. If the founder is likely to sell when they get a $500M acquisition offer, but the business might have IPO’d at a $5B valuation, then I might be picking a winning company but still fail at my portfolio strategy.
I want to understand if the startup is capital efficient - if not, it will increase my dilution in future funding rounds, which will substantially decrease my profit multiple upon exit. There are three types of costs that can hurt down the road but are not usually acknowledged by seed startups as a factor in valuation:
High total customer acquisition costs (not CAC) necessary to break in to a market with big moats.
Net working capital needs - when the working capital needed for operations scales linearly with revenue (perhaps due to payment terms with suppliers vs customers), and they can’t get it through debt financing, Buffett is famous for loving insurance companies for their negative net working capital structure (i.e. float).
Capex costs (hi rocket startups!)
In summary, no matter how attractive the product and the team, I always want to know whether the expected value is big enough to justify the risk and illiquidity that comes with startup investing.
Next week I’m going to dive into the world of syndicates (AngelList, etc) - how they work, and how it’s evolving.