The valuation of my portfolio

I’ve been asked off-board about my thoughts on valuation, specifically relating to Upstart, so here goes. I’m not going to dwell on P/S or any other ratio but rather on DCF.

Applying DCF to my portfolio

Given how difficult it is to do a proper DCF valuation and also that it is bound, by its very nature, to be incorrect to a greater or lesser extent in hindsight, what to do? Do nothing? Don’t attempt a valuation at all? Just look at P/S, which as Saul pointed out tells us nothing in isolation?

No, I vowed never to pay only lip-service to valuation after the financial crisis (I lost my shirt due to believing the EMH - but that’s a different story).

So I try. I go forth and come up with what I believe are reasonable revenue growth rates and FCF rates going 15 years into the future for each of my companies, and I discount that to today using a discount rate based loosely on WACC (10%) and I see where that gets me.

I don’t over complicate this exercise as it can be an enormous waste of time. I give it a bash using broad strokes. I put in what I think their revenue growth will be for about 5 years into the future and then I taper all of the companies’ growth rates down to the 20% and eventually 10%’s over time. And yes, of course this will turn out wrong. But this is a sense-check, not what drives my decision to invest or not. That decision is based on all of the good stuff about the company’s current performance a la Saul.

However given that I run a concentrated portfolio, all companies need to pass very strict hurdles, and this is just one more.

So, where did this exercise get me?

Here’s the interesting part for me, especially in the context of lots of noise about our companies being overvalued:

Every single one of the companies in my portfolio is currently undervalued based on my DCF calcs.

If they weren’t why would I buy them in the first place?? And this is after increasing the discount rate to take into account an expected interest rate hike. Inflation is irrelevant for this type of exercise as you are forecasting/guessing in today’s money and our companies will be able to adjust their prices in an inflationary environment imho.

And yes, this undoubtedly is open to debate as clearly the assumptions that go into the models are my own, but hey that’s the point right? By my reckoning all of the companies in my portfolio are at least 3x undervalued, which gels nicely with Saul’s goal of looking for companies that can 3x in the next years.

Here’s my list, from most undervalued to least undervalued (but with the least undervalued still about 3x too low by my calc).

  1. UPST
  2. PUBM
  3. MNDY
  4. AMPL
  5. SNOW (yes, SNOW. throw a couple of years of growth in the high 80%’s in there and see where that gets this company. Of course you need to believe that, though)
  6. S
  7. ZI
  8. SC
  9. DDOG

Portfolio weightings and valuations

So, why don’t I go all in, gung ho and just put it all in Upstart? My DCF valuation says that they are the most undervalued after all!

Because I could be sooo wrong. Especially on Upstart the likelihood of me being wrong is higher than the other companies (read the tons of great analysis on why this is more so for Upstart than the others on this list on the board - Bear has put it quite nicely I thought).

And I’ve been wrong many times before and will likely be again.

Because I have a lower likelihood of being wrong in my expectations of the SaaS companies with their predictable land and expand business models, NRR above 130%, operating leverage, customer additions, the SaaS companies on that list have a much higher weighting in my portfolio than the non-SaaS ones (and Snowflake is one of those non-SaaS ones).

Because they are much more predictable. I am much, much clearer in my mind about the likely growth path of SentinelOne for the next 3 years than for, say Snowflake or Upstart. For SentinelOne as good a proxy as any is to simply apply what happened to Crowdstrike three years ago to SentinelOne as the forecast in the DCF. Much more likely to happen than any forecast I can come up with for Upstart.

And why do I have a higher allocation than some in Upstart? Because I think my forecast could very well be right, and if I’m right, my Upstart shares will 10x in the next 3 years.


  1. should read ZS - ZScaler. Not SC.

I would be interested to see how you calculated your models and what numbers you came up with.

do you also have a way to validate your model as well?

WSM that was a very (!) interesting and useful exercise-- thanks heaps for posting it! It is yet ANOTHER approach to stress testing the Saul approach. Excellent!

Interesting exercise.
You said that you went just 5 years out and then I guess there’s the terminal value. How did you calculate the TV for these companies? The reason why I ask is that you could either apply a simple multiple to the n+1 FCF (and in that case you would pretty much fall back into the fallacies of the EV/S multiple, in a sense, given the large weight that the TV generally has on these companies’ DCF valuation), or you could figure out a growth rate in perpetuity, with the caveat that, should such perpetual growth rate be higher than the cost of capital that you’re using (and with a 10% wacc, there’s a high probability that that’s the case for these companies), the formula wouldn’t work.
Also, why 10%? I would use at least 15% for two reasons:

  1. that’s my hurdle rate in terms of the minimum required return I’m looking for (yours might very well be 10%, granted);
  2. a higher rate would factor in a higher risk, which is probably appropriate for these hyper growth companies.
    Thank you.

Posting to the board as others may have had the same questions:

  • I forecast for 15 years (as I stated), then do a terminal value using a perpetual growth rate of 3%.
  • My FCF% gets to between 20% and 35% for all the SaaS companies over time
  • I taper all companies’ growth rate down to at most 10% by the end of 15 years
  • Hyper-growth (>40% growth) lasts between 3 and 6 years and then tapers down
  • All of the companies end up as very large companies - compound interest is a force of nature, so you have to believe that they can get there
  • The discount rate of 10% is based on the WACC as I stated. That is the firm’s cost of capital. It’s got nothing to do with me (…
    ) It’s not my hurdle rate, or my cost of capital (which are two different things too). My hurdle rate, or required return for my portfolio is implicitly much higher. I want to 3x my investment in 3 years, which if you work it out is about 45% pa. But I don’t quantify a “hurdle rate” like that for myself; I just try to maximise my returns.