Conversation

3) At its heart, I've been trying to understand the following question: Let's say a company had $100m of revenue and $90m of expenses in the past year. It goes out to raise $50m. What valuation does the raise happen at?

4) Of course you need more information to know! For one, you need to know what the relationship is between that revenue and those expenses. In particular, let's say the business grows 10x, so next year it pulls in $1B of revenue. Do you expect $90m or $900m of expenses?

5) Well, it depends. Say it's a crypto exchange. If the expenses were "AWS + marketing + salaries", probably they don't scale linearly with revenue--maybe $1B revenue <> $200m of expenses; you don't have to scale up all the costs just because people are trading larger volume.

6) Say that instead, your "revenue" was taker fees, and your "expenses" came from maker rebates. your fee structure: 10bps taker, 9bp maker rebate. If your userbase/volume 10x, you'll make 10x the taker fees, but you'll pay 10x the maker rebates--$1B revenue <> $900m expenses.

7) So, in theory, you should calculate something like "net revenue" -- raw revenue minus expenses that scale with revenue. But the thing is, it's often hard to tell which expenses are which -- _especially_ if you're an investor and not the company itself!

8) So we run into Weird Thing #1: often companies will disguise scaling expenses as fixed expenses, and then VCs will invest in them at way too high of a valuation, thinking they have a business model that scales well!

9) People often quote "valuation ~ 20x": but 20x what? The company would either raise at $20b, $18b, or $2b, depending on how the expenses scale.

10) Ok, let's move on to another oddity. Take Company A: a crypto exchange that made $100m of revenue last year with no expenses. Say it's a "meh" exchange and isn't expected to grow or shrink going forward. Cool, $2B valuation?

11) Well, ok, you probably assumed it made all its money from exchange fees, right? What if it charges exchange fees in BTC. It made 2,000 BTC in profit last year, which is $100m. Fine, $2B still sounds ok.

10:07 AM · Oct 13, 2021Twitter Web App

12) Now, let's say that BTC went up during the year, so its average price was $10k but its ending price was $50k. The exchange still made 2k BTC, but.... marked to the time it made those BTC, it was only $20m. Which then appreciated to $100m. When should you mark the BTC to?

Replying to

13) Or, to take this further: the exchange fees were $0, because there is no exchange. But it started the year with 2k BTC, and those 2k BTC were worth $100m more than at the start of the year ($10m --> $110m). What's _that_ company worth?

14) Well, there is no company, really, other than BTC. It's just a company that owns 2k BTC. And that 2k BTC is worth $110m--you could replicate it on FTX if you wanted! So the company is worth $110m. This gets to an odd question: what makes revenue "recurring" vs "one-off"?

15) Revenue that's expected to recur should get a large mult (20x?); revenue that doesn't should just count towards the existing balance sheet. So you have Weird Trick #2: find some way to make one-off revenue but make it look recurring, and sell the company for 20x it!

16) And this gets to a weird question: Who made more this year: Coinbase, or Avalanche? Well, on the one hand, Coinbase made ~$6B or so, and the AVAX token made ~$12B (more fully diluted). But on the other hand, maybe that's not comparing apples to apples.

17) See, Coinbase made $6B in "recurring" revenue. AVAX's market cap went up $12B. But Coinbase's market cap went up ~$50B! So what's the right comparison? Well, one key thing: That $6B is Coinbase's and has been fully made.

18) The $12B from AVAX is only sustained if AVAX continues to perform well. Kinda like Coinbase's market cap. So I think that it should be compared to $50B, not $6B.

19) But, responds Avalanche, what if we kept doing what we did last year -- kept minting tokens? And the problem is: doing that would devalue AVAX and those future tokens.

20) And, Weird Trick #3: Projections. If a company is new, its revenue probably isn't mature. So how do you value it, if you expect growth? Well, you project out its future revenue, of course!

21) Where, here, "you" means "the company", and "project" means "create an excel spreadsheet with large growing numbers". You'd be surprised how many companies are already valued off of their 2025 revenue!

Quote Tweet

Jan 11, 2021

1) BAKKT is about to go public. There are a number of "interesting" things in their deck. sec.gov/Archives/edgar

Show this thread

22) Alright, so back to the original question: How are raise valuations _actually_ determined? In some sense, you take 20 * (projected annual revenue - projected annual expenses). But there's as much art as science in determining those.

23) And so, in practice, you see a tug of war. The company paints itself in the best light it can, making liberal use of dubiously recurring revenue and 2025 projections. The investors, in turn... absolutely shit on the company.

Beware Isolated Demands For Rigor

24) They have to disguise it, because they're supposed to be aligned. But they often do what they can: making unreasonably conservative assumptions, highlighting flaws, making isolated demands for rigor (slatestarcodex.com/2014/08/14/bew), and negging.

slatestarcodex.com

I. From Identity, Personal Identity, and the Self by John Perry:”There is something about practical things that knocks us off our philosophical high horses. Perhaps Heraclitus really thought …

25) Of course, as soon as the investment closes, their tune changes, and your company is grossly undervalued according to them; they're now _actually_ on your side! Or maybe, none of this matters. Maybe each VC firm *has* to invest in at least one crypto company this year.