Description

We just realized there is a rule that bans VIC users from submiting ideas that have been posted in the previous sixt months. We hope VIC community and admin forgives us. We also think this write-up adds value. But that is the reason why we posted it as G-Squared. This post will be brief. Nothing like the one we just posted on Meidong’s this week because.(1) Square is a household name and we do not want to repeat much about its model vs. what it has been already said in other write-ups (the last one as recently as March by inversionesinteresantes), and (2) we have tons of work and time constraints as our VIC membership deadline is getting closer. We hope to add some value by shedding light on some underappreciated aspects of the business. 

Square is a terrific business which reports confusingly, which we think adds to the general feeling of “great company overvalued stock”. We think it is cheap as we project 17%+ 5Y IRR in a base case (Square not closing the loop between its two ecosystems).

Why the reporting mess? Well... some of it is not SQ’s fault. SQ used to report consolidated “Adjusted Revenue” till Q3 2019 when the SEC disallowed it and required the company to use GAAP revenue instead. Adjusted revenue subtracts from GAAP revenue transaction costs, bitcoin revenue and deferred revenue. While this accounting change distorts time-series and peer cross-sectional gross margin comparisons and makes SQ look optically cheaper on a revenue-basis multiple, frankly it is not big deal. The only thing that matters, gross dollars, remain the same.

This is just the tip of the iceberg though and most of the reporting mess is SQ self-inflicted, burying its true underlying economics and growth potential. Since Q2 2020 the company breaks up financials for Seller and Cash App. While the split provides more clarity to investors, SQ still misclassifies its activities within each ecosystem.

Cash app’s reporting makes zero sense to us from an investor standpoint (probably more from a competitive lens). For one, SQ includes bitcoin revenue, which in 2020 accounted for more than three-quarters of revenue at virtually no margin. To put things in context Cash App’s reported 20% gross margins in 2020. This a bit annoying, especially for lazy investors who do not make the simplest adjustments (Spain + summer + reopening makes us good candidates to become such type), but the good news is that SQ also reports Bitcoin-related costs, so it is not that difficult for us to derive 80%+ Cash App margins ex-bitcoin with our mojitos-in-da-beach.

We decided however to get back into the office when realized SQ treats Cash Card, Boost (SQ’s reward program for Cash Card users) and Instant Transfer, which we estimate together comprise 80%+ of Cash App’s 2020 revenues and profits, as subscription-based instead of transactional. We took a quick survey in the beach and not even those who had already did five mojitos thought the business around the activity of swiping a card had something to do with their Netflix’s subscriptions, but for some mystery reason SQ does. To compound on the confusion since Q4 2020 SQ breaks down its GPV between Seller and Cash App ($103.7B and $8.5B in 2020 respectively for a $112.2B total). The thing is that as TPV is a transactional item the $8.5B figure excludes Cash Card, Boost & Instant Transfer as SQ considers them subscription based.

What do those $8.5B refer to? What is Instant Transfer and Cash Card’s actual TPV? Cash App’s “actual TPV” = P2P + P2P Credit Card funded + Cash Card + Boost + P2Businesses + Instant Transfer + Bitcoin + Stock Investing. We leave out of the equation however Bitcoin and Stocks – they add to GPV and generate revenue but virtually no profit – and P2P as it does not monetize to keep it a frictionless product that boosts traffic from existing users and serves as Cash App’s main acquisition funnel.

Therefore, Cash App’s “profit-generative TPV” = P2P Credit Card funded + Cash Card + Boost + P2Businesses + Instant Transfer. The $8.5B GPV is only for P2P Credit Card and Payments to Businesses. Heck! No sane mine doubts $8.5B is a respectable number. However, it looks relatively unimportant as we realized the other components of the equation generate $76B TPV. You might wonder how we back these numbers if SQ omits them. Let’s start with Cash App excluding Cash Card.

We investors gather, process, and analyse data, looking for supportive and disconfirming evidence to fill holes in our base assumptions, also hoping to a have a bit of luck. No matter how hard we work though, every investment thesis ultimately has an embedded leap of faith to it. Cash App’s 130 MAUs by 2025 for the case. You might wonder how this number makes sense given just 20% of the US population, or 66 million, are unbanked/underbaked, precisely Cash App’s main user base. Frankly speaking, we do believe the 64 million gap is not a stretch. For one, Cash App’s coolness and social enticement means the 11m Americans neither-unbanked-nor-underbaked aged 15 to 34 are addressable.

Two, we believe Cash App’s tentacles will extend far beyond its core user base. The fact that most of Venmo’s 54 million users have access to a bank account seriously weakens the hypothesis that a bank-independent wallet is merely an underbaking phenomenon. On top of this many Americans have turned their interest from Venmo to Cash App. One explanation for this is unlike Cash App, which partners with influencers and is active in social media, Venmo’s marketing is nearly inexistent. More importantly Venmo lags Cash App’s innovation pace. Never mind Venmo launched its P2P network in 2009 and Cash App four years later, Cash App got ahead in every feature on top of the initial wallet. For example SQ launched its debit card (Cash Card) in 2017 and Venmo took till next year to do so. SQ pushed Boost, Direct Deposit and Bitcoin trading in 2018 while Venmo catapulted Boost in 2019, and Direct Deposit and Bitcoin both in 2020. Cash App’s innovation lead and its consequent product superiority manifests on its thunder-speed adoption. Google Trends does a pretty good job in showing the evolution of each network in customers’ top of mind.  

While Cash App and Venmo might coexist for years we believe network-effect dynamics will unbalance the fight in Cash App’s favour. This is not to say Venmo will disappear. In fact it might have an even larger “user” base in the future. More relevant than the # of users itself is their level of engagement within the app as it is the main contributing factor to attract monetizable TPV. All in all we believe 130m Cash App users in the US alone does not signal mental insanity on our side. In fact just think about end-game adoption in the US for other widespread best-in-class services. Amazon Prime has 148m subscribers (paying subscribers!), leave Facebook 200m+ users aside.

If you still feel we are too bullish on MAUs you might add the optionality of Cash App successfully scaling in some international markets. In this regard we just saw on LinkedIn in this regard the product operations lead at Revolut, a leading UK-based fintech and certainly a Cash App competitor in the country, has just moved to Cash App. The challenge for Cash App to be successful in Europe – and by extension in our country, Spain – is that there is not a meaningful amount of people who are either unbanked or underbanked, which begs the question whether in the absence of early adopters who really need the product the product quality itself will do the work. We found two antipodean situations to judge this but no concluding evidence. In the US Venmo scaled and not precisely out of the unbanked, but backing out of this that a P2P network might work out just of banked customers might be wrong. This is because Venmo launched in 2009 and there was not comparable bank-owned P2P by the time until Zelle came out with its basic P2P solution in 2017.

In fact, the Spanish case suggests the opposite. Verse – acquired by SQ in 2020 – and Bizum (the bank-owned solution) launched both in 2016. While Bizum now has millions of active users Verse has just thousands of them. Concluding however out of it that banks unequivocally side with victory as soon as they launch a P2P MVP early might be also mistaken. This is because Verse is also a plain P2P MVP (it just launched its debit card in 2020) so we can’t conclude that product superiority by itself eats the bank-owned solution for breakfast. We intuitively have sympathy for this best-product-wins view though. Why? When we met Verse’s CEO we learnt he (1) might be more than half an hour late to a meeting (2) does not apologize for it (3) lacks a vision and mission for the company (4) spent the meeting repeatedly asking which stocks we owned so he could buy them, (5) in case SQ fires him (said literally) or at the end of his four-year commitment tenure he plans to launch a SPAC of whatever bullshit trendy thing that might screw investors but makes him rich (Chamath’s Spanish version), (6) prides himself for selling Verse to SQ for around $100m (we did not ask, he just came up with it) and (7) Cash App’s lead Brian Grassadonia contacted him first expressing interest in Verse. While $100 might not sound much for SQ scale, Brian is a smart and visionary dude unwilling to throw that amount of money out of the window in some fancy venture run by an incompetent. He might have seen something else. The most plausible logic for the acquisition is that upon implementation of SQ’s architecture and features Verse’s value proposition will be a step-function of the incumbent Bizum solution, inevitably scaling up.

The second assumption you need to believe to be ok with our projections is that the instant transfer fees do not get competed away. While we think charging 150 bps for the service might look excessive, Zelle offers free instant transfers and nothing has changed. We wonder what would happen if Venmo decided to remove these fees but we find comfortability in a Disney analogy. Disney’s parks are such differentiated experiences from other recreational parks that visitors do not care much of paying $5 for a coke even if they know they can get the soda for a fraction of that price elsewhere. We believe Cash App’s ease of use and high engagement levels make the service so unique that users see the product in a holistic way instead of evaluating the merits of each of its constituents separately (price included).

Now let’s turn to Cash Card projections.

Two main assumptions for cash card. First, the 55m cash card users by 2025 are in turn a function of (1) the number of cash app users (already explained) and (2) the adoption rate, which we see following a steady upward trend as Cash App increases its functionalities, giving a reason to potential users not to pay more at a bank for the same services. This should mean more users convert the platform into their primary account, which is the main driver for cash card adoption.

The second assumption is that the annual spend per cash card user mean-reverts to debit card consumption per capita nationwide levels. The logic behind this trend is (1) incremental users will be wealthier than early adopters (they were unbanked /underbanked, right?) so they have the ability to spend more and (2) as Cash App gains primary-account status amongst more users spend per user will inevitably rise.

Cash Card’s reward program Boost deserves a mention now. Although we classified it within Cash App ex-cash card to provide more clarity, Boost is inherently linked to cash card transactions. Boost depresses cash app’s economics as SQ, not the merchants, currently funds the rewards for customers in an attempt to accelerate adoption. It has the potential to be a huge business though. How large? The listed-peer Cardlytics might serve well for the case.

Cardlytics operates a platform that connects marketers, financial institutions, and customers via cash back rewards that show up on customers’ banking apps and webs so they can use them when swiping their cards. Let’s say a marketer like Starbucks wants to spend $15 per customer to attract store traffic. In such transaction $5 accrue to Cardlytics, $5 to the banking partner and $5 to the customer in cashbacks. While Cardlytics exhibits some desirable traits as a business Boost offers a lot more to like. Let’s leave for a moment Cardlytics is run by a bunch of former bankers who were useful in the early days to develop the banking partnerships but are clueless on marketing and technology. OK? Let’s think Messrs. Dorsey and Grassadonia manage Cardlytics. Even if that was the case Boost benefits from three structural advantages. One, Boost does not need a banking partner in the equation as it connects marketers and customers directly through cash app. So for every $15 a marketer spends in Boost, as the pie splits between 2 actors instead of 3, customers get larger discounts and Boost makes more money. Two, rewards are more effective in Boost by orders of magnitude, becoming the platform of choice for marketers. Effectiveness is a function of customer engagement with the platform and thus frequency of use. Some might argue Cardlytics has a larger-user base advantage (all the bank’s customers) but (1) that gap will diminish over time and (2) each cash card user is equivalent to bunch of Cardlytics users as connecting to a bank app is a dreadful experience so people do not pay much attention and rush out of it. The third advantage is a direct benefit of having two interconnected ecosystems. SQ has POS. Cardlytics and its banking partners don’t. While for a small one-item transaction might not matter much (a $2 from a coffee shop can’t be something rather than a coffee), for large ticket multi-item purchases makes all the difference. When you buy $70 in groceries at the local supermarket all info the bank gets (and thus Cardlytics) is a bank account entry with the name of the merchant and the $70 charge. But now look at the picture below. Through its grocery-based POS SQ get granular information on the products/brands/quantities that comprise that $70 bill, conferring a far more interesting value proposition for marketers in terms of targeting.

We projected $5.1B Boost revenue by 2025 Cardlytics currently sells for 8x 2025 consensus gross profit. Applying the same multiple to Boost – unadjusted for the far superior business it is – we get $40B for boost standalone, 40%+ of SQ’s current market cap. How do we get that $5.1B revenue? First, we estimate that ~20% of the spend per cash card user flows through, which is reasonable given the advantages we already described plus the locational nature of the product (the capacity of the platform to connect prospective merchants around the places where customers spend most of their time, i.e. where they work, live, etc). Second, we estimate a 7% take rate. To evaluate this just think of the spectrum of demand generative platforms (the Estys, Amazons and Facebooks of the world) and how much they charge. You will quickly realize that our assumption is in the low end of the range with arguably better targeting, as no other demand generation platform can provide actual 1P data that rather than intent or interests actually comprises an actual purchase.  

Regarding cash App economics Marqeta’s S-1 is very helpful. The main take always of it are that Marqeta’s contract with Square ends up in 2024 and that Square comprises 70% of Marqeta’s revenue. It may not surprise you that the contract will be renegotiated dramatically to SQ benefit.

The sun is almost rising here in Spain, and with less than 20 min to meet our submission deadline we will not extend much about Seller. The seller vertical SQ allocates for good reason its point-of-sale transactions (either a physical payment terminal or an online solution) to the “transaction-based” basket, the services to help merchant to run their businesses – website hosting, domain registration and tools for customer engagement, billings, and employee management & payroll – to “subscription and services” and the roll-out of its physical POS devices as “hardware”. Instant transfers and loans to sellers (the latter aka SQ capital) are however assigned to subscription when are transactional in nature, which is important because the former typically are 95%+ margin items but the latter command lower ones. While it is no big deal either as SQ capital loan application, acceptance and underwriting are scalable online processes with not much overhead and collection comes through a cut of every subsequent transaction in the merchant POS it stills dents gross margins a bit. Here find our projections, we can further discuss them on Q&A.

Catalyst

- Cash App user and engagement growth 

- POS roll-out