Conversation

SBF
@SBF_FTX
Replying to
@SBF_FTX
3) The problem is that they report revenue by "retail" vs "institutional". But the *real* split is by platform, not user. So, yes, their retail take rate went up from 1.10% to 1.23%. At the same time, their institutional take rate went down, from 0.029% to 0.025%.
4
28
SBF
@SBF_FTX
4) Coinbase's *real* split comes when you look at the mobile app (and coinbase.com) vs pro.coinbase.com. App/Coinbase charge ~3%; pro charges ~0.02%. (150x higher on the app!) Retail *correlates* with mobile app, but not perfectly.
pro.coinbase.com
Coinbase Pro | Cryptocurrency Exchange
US-based crypto exchange. Trade Bitcoin (BTC), Ethereum (ETH), and more for USD, EUR, and GBP. Support for FIX API and REST API. Easily deposit funds via Coinbase, bank transfer, wire transfer, or...
3
28
SBF
@SBF_FTX
5) So maybe their take rates per user went up, and yes, their userbase did seem to get more retail heavy. But say, for instance, that Coinbase finds an individual whale that trades $10m/day, and onboards them to pro.coinbase.com.
pro.coinbase.com
Coinbase Pro | Cryptocurrency Exchange
US-based crypto exchange. Trade Bitcoin (BTC), Ethereum (ETH), and more for USD, EUR, and GBP. Support for FIX API and REST API. Easily deposit funds via Coinbase, bank transfer, wire transfer, or...
2
17
SBF
@SBF_FTX
7) Then, 3 months later, that user creates a single-member LLC and switches over. All of a sudden, Coinbase's reported retail take rate spikes way up! Because they offboarded a "retail" user on the relatively low-fee Coinbase Pro; or rather, re-classified their volume as insto.
2
22
SBF
@SBF_FTX
8) They show volume and revenue splits by individual/institution, but the more interesting split is by coinbase vs coinbase pro, because that's where the fees really differ. So apparent moves between retail and insto can be misleading!
3
21
SBF
@SBF_FTX
9) Anyway, the second thought I had on
@fintechfrank
's analysis--about OTC and prime. Yes, those did grow. But they're almost certainly under institutional transaction revenue. That whole category was only 4% of Coinbase's revenue, so it can't have been a big factor!
2
19
SBF
@SBF_FTX
10) Remember, Coinbase's whole business model is mobile app (vs pro) fees. So volume beat expectations because of OTC, but it didn't lead to much more revenue, so it didn't cause the earnings beat.
6
30
SBF
@SBF_FTX
11) And the non-transaction revenue, coming in at $213m, or about 10% of net revenue in Q4? Well, half that comes from "blockchain rewards", i.e. staking.
6
26
SBF
@SBF_FTX
12) And, as always... "...transaction expense growth, which includes staking rewards paid to users." Coinbase calls staking it net revenue, but it's really gross revenue. Customers stake tokens and get yield; Coinbase books that yield as revenue and expenses.
11
52
SBF
@SBF_FTX
13) The real breakdown, roughly, of true net rev: Mobile/coinbase: $2b, ~88% Retail on Pro: $100m, ~4% Insto on Pro: $75m, ~3% Insto on OTC: $25m, ~1% Custody: $50m, ~2% Other: $50m, ~2% (Backing out some guesses here, could be off by a bit)
22
94