Proxy Valuation Modeling

Balance what you have with what you need

Sorry for the late post this week - my wife and I were both very sick for the first half of the week but are much better now.

I have been reading a super interesting book called Investing: the Last Liberal Art, very slowly over the last year. I'd recommend it to anyone who wants a slightly different and very thoughtful take on the mental models used in investing. The book begins with a discussion on Charlie Munger's concept of a lattice work of mental models and the value in using a variety of perspectives and then seeing how they coalesce around a certain thesis or trade. The author, Robert Hagstrom, then goes through physics, biology, sociology, psychology, philosophy, literature, and math, bringing some of the key high level concepts in each discipline into focus from an investing perspective. I am still in the section on math, but there are a couple paragraphs in the philosophy section that jumped out at me with regard to venture investing and using higher order models to improve the accuracy of long term value assessment.

Excerpt from page 102 (my emphasis):

What is the best measure of value? Most believe John Burr Williams's theory of discounted cash flow (DCF) is the best model for determining economic value. We should think of Williams’s DCF as being a "first-order model." However, many investors shy away from its inherent difficulties. Instead, they drop down one level of explanation and select one of the second-order models – perhaps low price-to-earnings ratios or some other accounting factor-based measure – which they rigidly hold up as the only correct approach.

The stock market is a giant discounting mechanism that is constantly repricing stocks. There are occasions when the stocks that offer the greatest discount to the company's cash flows (the DCF model) are stocks with low price-to-earnings ratios; at other times the greatest discounts can be found in those stocks with high price-to-earnings ratios. No one metric is absolute; none is always right. Pragmatic investors can, and should, apply any second-order model that is fruitful and discard any that are worthless, all without violating the first order.

Remember, James tells us that even "the most violent revolutions in an individual's beliefs leave most of his old order standing." Even when we adopt a new idea, we can still preserve the older ones with minimum modification. From a pragmatist's point of view, it is permissible, even advisable, to search for those explanations that work. "Stretch them enough to make them admit the novelty," James said, "but conceiving in ways as familiar as the case leaves possible.”

The philosophic foundation of successful investors is twofold. First, they quickly recognize the difference between first- and second-order models, and as such they never become a prisoner of the second-order absolutes. Second, they carry their pragmatic investigations far from the field of finance and economics. It can be best thought of as a Rubik’s Cube approach to investing. The successful investor should enthusiastically examine every issue from every possible angle, from every possible discipline, to get the best possible description – or redescription – of what is going on. Only then is an investor in a position to accurately explain.

I wholeheartedly agree that the DCF is the only true first-order valuation model. All other perceived values are based on the long term ability of a business to generate cash flows. But as Hagstrom astutely points out, this isn't always the case because of interim market forces, and more importantly in the world of venture, this is almost never the case because our precision calculating cash flows and discount rates is so bad at the early stage. As a result, we must look to higher order models and use them as proxies in order to help us understand the fundamental value of the business. Hagstrom makes another point which I wholeheartedly agree with: "Pragmatic investors can, and should, apply any second-order model that is fruitful and discard any that are worthless, all without violating the first order." It is critical to constantly create and evaluate the second (and higher) order models based on the data available and the specific situation, but it is even more critical to not violate the first order model. It's also important to note that fruitful in this case is hard to define given the investment horizon and variety of potential outcomes.

For me, the key is to focus on models and variables where we have the best quality data (highest precision) and use those as a proxy for a long term value assessment without forgetting the concepts behind the primary principles of value, always ensuring that we can map back to long term cash flow generation viability (i.e. unit economics done right). Put another way, instead of replacing the DCF entirely, use the higher order models as stepping stones to get from what we know now to what a DCF could look like many years down the line.

For a typical high velocity software business, it might look like this:

  • First order: DCF

  • Second order: Net Income

  • Third order: EBITDA or FCF

  • Fourth order: Revenue or ARR

  • Fifth order: Revenue Growth or ARR Growth

  • And then (roughly in order): Unit Economics; Engagement Metrics; Users Metrics; Signups; Waitlist

  • If we step into the more qualitative world, these metrics become signals: Who else is investing; Where did the founders work/school; How much have they raised; What is the competition

As you look at earlier companies, you must rely increasingly on the higher order metrics, and it takes significantly more foresight and experience to ensure that they can realistically map back to a long term value-creating, cash-generating business. As I have explained to many friends, it's not that a waitlist is a wrong way of looking at valuing a business, but it's certainly not sufficient as a standalone metric.

One of the exercises I go through with my companies early on is to help figure out and optimize the longer term viability of the business model. I find that this exercise can be done in a matter of hours and can inform almost every decision a founder makes from then on - hiring, pricing, product features, positioning, gtm, marketing, and more. In a world with new and innovative business models, the risk of not doing this exercise is a long-term unit economics-negative and cash burning business. If that is the case, a company may be able to sneak out a strategic exit, but it's almost like musical chairs - where will you be when the music (i.e. funding) stops?

As always, I am open for discussion (@alexoppenheimer on Twitter) and will hopefully be able to share one of those unit economics and business viability exercises in detail in the future.

Also, this is a post I wrote a while back on the subject as it relates to pubic SaaS valuations: