Growing up, I wasn’t popular in the conventional sense (and by the conventional sense I mean having friends and such). As a result, I spent a lot of time reading. Mostly it was epically nerdy science fiction (Heinlein, Piers Anthony, etc) but there were a few conventional writers I liked as well, embarrassingly including Judy Blume. Somewhere in between her kid stuff and her adult stuff (if you never discovered WIFEY I feel sorry for your loss) was the adolescent classic TIGER EYES. While it now seems hard to imagine YA fiction without a dystopian future plot, trust me, this book was good. Two salient points: a)my now wife was the model for the cover art on the book (!); b)there is a reoccurring quote that I’m using as the title of this blog, “Comparisons are odious”.
I’ve been running “comps” for 20 years now, since I started at Bain Capital in 1995. For the uninitiated, comps are a list of similar (specifically: comparable) companies that are theoretically useful in determining the proper valuation ranges for an investment you are thinking of making. Eg, “Yeah, I know paying 10 times revenue seems high, but Workday trades at 14X.” Or, “How can you pay 10 times revenue for Tesla when Ford trades at 10X earnings?!” I’ve never really understood the logic of this type of analysis, and I’ve actually become convinced it’s the diametrically wrong guide to use as an investors. Given that we hold our investments on average for 5-7 years, shouldn’t we do the opposite of what Mr. Market is doing at any given moment? Comps are a popularity contest and rarely durable.
If at any given moment, horizontal saas companies trade at a premium to vertical saas companies, what are we to make of that? Presumably investors believe that the TAM (total addressable market) benefits of broadly applicable software trump the sales and marketing inefficiencies. Suppose it becomes clear that, at scale, vertical saas companies have 500 basis points of additional margin due to structurally lower go-to-market costs … don’t we think revenue multiples will expand, on average? Surely we can anticipate changes in the conventional wisdom of this sort over a multi-year hold period. (If you have any doubt about the likelihood of this, let me remind you that cardiologists have recently reversed their position on cholesterol and are recommending people eat eggs every day. To my long list of regrets I now add every single egg white omelette I’ve ever had.)
If anything, low multiple comps can be a draw to a segment, in that they usefully point out that the incumbents are doing something fundamentally wrong, to wit:
- Lead gen companies typically trade at low multiples because their customers (or users) are transitory and bounties fluctuate cyclically. Credit Karma flipped this on its head, in that they have loyal users who trust them to recommend, in context, a broad variety of pre-approved financial products. How are the lead gen comps relevant to Credit Karma?
- Issuer processing companies trade at 15-16X EBITDA (a high multiple, by the way, on a relative basis). Google bought TxVia, a prepaid issuer processor, for [A BIG NUMBER] X Revenue… TxVia and Google both saw how mobile wallets were creating a new use case of account based consumer financial services that had nothing to do with traditional prepaid.
- Debt collection service providers trade at poor multiples historically, as befits a cyclical commodity service with regulatory overhang. To an entrepreneur like Ohad at TrueAccord, that’s the opportunity … he can do it better, and flip the model on its head. Should we value TrueAccord in the context of the traditional comps?
The relevant thing about comps, ultimately, is not the revenue or EBITDA multiple levels themselves, but the implicit logic behind them. If an entrepreneur can build a company that challenges the traditional economic logic of a segment, they are likely on to something big.