Description

Let’s start with the obvious: Square is not a cheap stock. That said, the business has been on fire operationally and it got me interested in exploring the bull case. It is probably not a very good risk-reward but I do think you can make the numbers work if you hold your nose a little bit.

One thing to note is that SQ doesn’t actually trade for 8x revenues (this is what shows up on TIKR and I believe CapIQ shows something similar). The true revenue base is what SQ used to call adjusted revenues. This is because some costs can scale over time (like non-BTC Cash App revenue) while others cannot like BTC costs and transaction-based costs (aka pass-through interchange costs). You have to account for interchange costs that are paid through to the banks but aren’t accounted for in SQ’s net revs like they are for most acquirers. Rather, those costs are under the “cost of revs” bucket, specifically “transaction-based costs.” 10K: “Transaction-based  costs  consist  primarily  of  interchange  and  assessment  fees,  processing  fees  and  bank  settlement  fees  paid  to  third-party  payment processors and financial institutions.”

SQ used to report adjusted revenues which backed out these transaction-based costs (for some reason the SEC now disallows them from reporting that). But the walk is: adjusted revenue = GAAP net revenue – transaction-based costs – bitcoin costs + deferred rev.

So it’s not really trading for 8x the true revenue base. TIKR shows consensus GAAP revenues of $13.2bn in 2021 (which would look like 8x revs) but that’d probably mean ~$5bn in adjusted revenues (this is based on the GS model’s walk from GAAP to net revs). So in reality, SQ is trading for more like 20x true net revenues. 

Is that still attractive? Now that we have the right revenue base, we can compare the valuation to other acquirers. The goal here is to do the kind of aspirational VC math on SQ by thinking through what margins might shake out to be in five or ten years from now (i.e., at scale).

Let’s first look at Square’s Seller business. This is their original business that sells Square’s custom point-of-sale equipment to merchants for payments processing and increasingly additional software services.

There is an incredible number of services offered through the Seller business and I would encourage interested readers to check out Square's website to see all of them. Their payroll offering, for example, looks fairly robust for the average SMB: https://squareup.com/us/en/payroll

In thinking about where this business’s margins can shake out, we can look at the EBITDA margins of comps like FIS, FISV, and GPN:

For Square’s Seller business, the comps are clearly all high margin. But they are also much much bigger in terms of payment volume processed. So can SQ Seller ever get there?

Here is the bull case: SQ Seller is really an omnichannel business as it has no indirect sales force (all word of mouth / direct customer acquisition) unlike the traditional acquirers who use ISOs to go pound the pavement and try to sell POS devices. Additionally, ~60% of gross profits are coming from software, SQ Capital, and Instant Transfer / Square Business debit card—none of which have an interchange component and should all be very high margin revenue streams in the end state. I think this is perhaps the most important point about Square’s Seller business that people miss. Financials analysts look at SQ’s valuation, scoff, and go long FIS/FISV (here’s an example). Meanwhile fintech investors also have a difficult time “squaring” SQ’s multiple versus other payment processor peers. What they are probably both missing is that Square’s Seller business is as much (or more) about its software and value-added services offerings than anything else. The payments angle is just a commodity that they lead with in order to drive much higher-LTV and recurring SaaS products. 

Over time, more and more of the Seller gross profit has come from these services rather than the commodity business of processing payments (sidecar payments):

There are 4-6x LTV/CACs in this business with ~3 quarter paybacks even though close to half of payment volume now coming from larger sellers with >$500k in annual volume. 

So yes I do not think it is crazy to think that SQ Seller EBITDA margins at scale should settle out between where FIS FISV GPN are (40-45%) and potentially Adyen at the high end at ~55% even if it doesn’t reach their payment processing scale. You have to marry the payments margins with the software margins. At the midpoint let’s say Seller can get to 45% EBITDA margins given two offsetting features: negative gross-margin hardware revenue and high-margin software revenues.

Now on to Square’s other business: Cash App

It is harder to figure out where Cash App margins will settle out but not too hard to guesstimate. There are three main revenue streams within Cash App:

• “Sending” equals revenue from Instant Deposit, Cash for Business, and P2P transactions funded with a credit card

• “Spending” equals revenue from interchange on Cash Card transactions, fees on ATM withdrawals, and interest on customer funds

• “Investing” equals bitcoin revenues. You can buy stocks on Cash App too but zero commissions so it’s just an additional use case meant to bring you there. Really nice that you can have Cash App be your Direct Deposit account and can also buy stocks and BTC right in there without having to also open a separate brokerage account.

Triangulating these two disclosures on Cash App helps us understand where unit economics are / can go:

It might be hard to read but the footnote in the screenshot above (from 2020 investor update pres) says “Investing equals bitcoin gross profit” even though the graph is showing revenue for the Sending and Spending streams. So we know that bitcoin revenue—despite it being a lot as the 10Q table shows—is actually a very small gross profit driver. Let’s call it 5-10% gross margin on $2.8bn in bitcoin revenues during 9M ended Sept 2020, or $140-280mm in gross profits from bitcoin against $850mm in overall Cash App gross profits. That leaves $570-710mm in non-Bitcoin Cash App gross profits against $981mm in non-Bitcoin Cash App revenues, implying a 58-72% gross margin for Cash App ex BTC purchases. That’s 65% gross margins at the midpoint, and maybe that turns into 30% EBITDA margins (conservative I think as this is an entirely network effects-driven and virality-based business). BTC-related EBITDA margins should basically be the gross margins so 7.5% at midpoint.

The problem, though, is that right now ~75% of Cash App revenues are coming from Bitcoin (mainly because this was such a strong Bitcoin buying year). If that holds, then consolidated EBITDA margins for Cash App would be ~13%. Of course, BTC is unlikely to be that big of a driver every year (e.g., in 2019 BTC was only ~45% of Cash App revenues rather than ~75% in 2020), and the goal is to continue to add functionality that has higher margins on it much like MELI is doing—more asset mgmt. products, lending / personal loans, BNPL, insurance, etc. But let’s say they can’t innovate any further and so consolidated Cash App EBITDA margins tap out at ~13%.

Then putting it altogether that’s EBITDA margins of 45% at Seller and 13% at Cash App. Based on the revenue split from 9M ending Sep 30 2020, the consolidated normalized margins would be 26%. Of course, Cash App is going to continue to grow faster than Seller for the foreseeable future (negative mix shift for margins), but at the same time the Bitcoin run in the last 9 months has also been insane which has juiced BTC’s revenue contribution to Cash App, so let’s say in the end state consolidated margins still end up at 26% on EBITDA.

Recall we said SQ is trading for ~20x true net revenues. 26% EBITDA margins on that base means SQ is trading for ~75x NTM normalized EBITDA now. So… not that appealing in isolation.

But what about the growth? SQ’s topline accelerated massively during the pandemic and I don’t think growth at that pace is likely. Even still, SQ grew the topline at an average of 50% annually from 2016 to 2019 (i.e., pre-COVID). Consensus still has it growing 40% in 2021 despite more than 100% growth in 2020 because payment volumes should pick up considerably in the Seller business as things reopen.

So to do some back-of-the-envelope valuation math, let’s say SQ continues to grow the topline at a 40% rate over the next 5 years but the valuation gets cut in half from ~75x NTM normalized EBITDA to 38x at exit. Let’s also assume that the meager FCF yield here is burned on M&A / R&D that doesn’t pan out. In that case, the rough IRR should be the topline growth less the effect of multiple compression on the IRR, or 40% less ~10% = ~30%. (Note: multiple compression’s effect on the IRR is calculated as 38 / 75 = 51% spread over 5 years, or ~10% per year; I also have not accounted for share count dilution due to SBC).

I am curious to get the views of others who have spent more time here. There are some high-conviction positions in SQ from smart investors: 20% position for Bares Capital, HMI’s largest position (11%), 10% for Shawspring, Dragoneer’s 5th largest position, 7th largest at Coatue (bigger than NFLX AMZN FB). But almost all of those guys got in way way earlier when SQ was at ~8x true net revenues, equating to ~40x normalized earnings based on our margin math above.

One major risk is fee compression in the Cash App business from Instant Deposit, which is the overwhelming majority of Cash App’s gross profit pool today. The biggest question is how that gets competed down over time, perhaps from the Fed’s proposed real-time payments network set to launch in 2025. That would allow all of us to send money to each other in real-time and thus render the Instant Deposit business useless. My initial thoughts here are:

• We already have real time payments in the US (link)! And no one uses it… we also have Zelle which is essentially a RTP network. Ultimately I believe these products have not taken off because they are intermediated / distributed by the banks, and banks like their high-margin interchange revenue. This is the same reason why V and MA’s position is so durable. Banks aren’t going to give up all their interchange revenue (that has no capital requirements on it) in order to move money around faster for all of us

• I do expect yields will compress in Cash App on the instant deposit stuff but they will just make it up on other products. They’re already doing that now with the Cash Card that is a debit card linked up to your Cash App account (SQ collects interchange revenue on that).

• More seasoned Cash App cohorts are a lot less reliant on Instant Deposit too:

Link to the cited 2020 business update pres in case helpful: https://s27.q4cdn.com/311240100/files/doc_downloads/Investor-Update-March-2020.pdf

Catalyst

This business is already catalyzed I'd say