some thoughts on Adyen

Before exploring the competitive concerns that have sent shares cratering nearly 60%, it’s worth reflecting on what it is that got people excited about Adyen in the first place.

In a typical online transaction, a shopper’s credit card details are picked up and encrypted by a gateway, then routed to the appropriate merchant acquirer. The merchant acquirer, through a payment processor, sends that information to Visa or Mastercard, who routes that information to the card issuing bank, who authorizes the merchant acquirer to accept payment. Merchant acquirers and payment processors are often seen as one in the same but they technically play different roles. Merchant acquirers are financial institutions. As members of card networks, they bear responsibility for chargebacks and disputes, ensure that the merchants they onboard comply with regulatory requirements and, assuming the transaction is approved, receive money from card issuing banks. Payment processors are software. They intermediate transaction details between the gateway and the acquiring bank.

The confusion between acquirers and processors is understandable as incumbent processors like First Data and Vantiv were once owned by banks before being cleaved off as separate entities, creating a daisy chain of interactions where most steps in the payments flow were handled by different entities. A gateway like CyberSource or Braintree was distinct from a merchant acquiring bank like Wells Fargo, Bank of America, or Citigroup, which in turn was separate from a merchant acquirer processor like First Data or Worldpay. The fragmented, arms-length arrangements forced merchants to maintain a host of connections. A large multinational retailer might have relationships with dozens of different merchant acquirers and gateways across the globe.

In the 2010s, payment service providers (PSPs) like Stripe emerged to simplify this process by sitting between the merchant and acquiring banks. Instead of opening accounts directly with acquiring banks, a process that could take weeks, a merchant could instead open an account with Stripe, who would rent the Bank Identification Number (BIN)¹ of acquiring banks (”rent-a-BIN”). This arrangement gave Stripe the right to onboard merchants according to the acquiring banks’ underwriting rules, provide ongoing account services to the merchant, authorize transactions, and collect transaction data. Stripe pays a fee to the acquiring bank, who still assume risk of loss but otherwise retains most of the merchant acquirer economics.

So, to a writer with a paywalled blog on Substack it might appear that Stripe is talking directly to card networks and sending funds from the subscriber’s bank to yours. But under the hood what’s happening is Stripe is sending encrypted card details to an acquiring bank like Wells Fargo or PNC, who in turn is using First Data or some other processor to interface with card rails. Some payment providers will say they offer “merchant acquiring services” or even refer to themselves as a merchant acquirer. But usually what they mean is they are licensing a BIN from third party financial institutions. It is crucial that PSPs process volumes locally, either by owning a bank license or renting one, in each country that they accept payments, as domestic transactions are settled faster, avoid FX fees that attach to cross-country processing, and more likely to be approved by local card issuing banks, who see purchases within their borders as less risky.

The card networks we’re most familiar with in the US are but a fraction of payment methods available globally. Bank transfers are used in 1/3 of online transactions in Germany and in just over half of online transactions in the Netherlands. In India and Thailand, bank transfers and wallets account for around 60% of e-commerce purchases. Moreover, a wide variety of card schemes, bank transfer apps, wallets, and other payment methods are used in different regions and countries, including MercadoPago in South and Central America, Samsung Pay in Asia Pac, Boleto and PIX in Brazil, iDEAL in the Netherlands, Alipay in Asia, PayPal in Germany and the US, etc. Often times, PSPs will rely on aggregators to access the most widely accepted payment methods in every country. For instance, Stripe, PayPal, and Global Payments integrate with an aggregator called PPRO, who in turn maintains connections with more than 400 non-card payment methods across the globe.

Adyen started as a gateway but took things several steps further. Instead of renting BINs or integrating with an aggregator, they own banking licenses outright wherever they can² and directly integrate with local payment methods. Combining the gateway with merchant acquiring and processing means Adyen can underwrite merchants according to their own compliance standards and save on fees they would otherwise pay a third party acquirer. But more importantly, it gives them greater visibility into and control over the payments flow. As an acquiring bank with direct access to card rails, Adyen can iteratively tweak messaging details in an authorization request to improve the chances that the issuer bank will accept it. If a payment fails due to a connection issue at the level of the acquiring bank or processor, they can more easily see and rapidly fix it than most other payment service providers, who obviously can’t inspect the systems of the acquiring banks they rent BINS from. Adyen, more so than any other PSP, is very opinionated about owning the full stack. They would rather not process payments in a country at all than gateway to third party acquirers, whose variable performance dilutes their exacting service standards.

Control not only applies to the payments flow but extends to the enabling technology. Adyen runs a single platform that it built from the ground up. A merchant need integrate just once to access processing capabilities anywhere in the world, both online and at physical point-of-sale. All merchants share the same platform. Upgrades and new products are rolled out across the globe at once. By contrast, legacy PSPs like Worldpay stitched together a patchwork of providers through a flurry of acquisitions. After struggling to build unified platforms, most legacy providers gave up and came to accept the disparate hodgepodge. First Data’s integration saga went on for about a decade before they threw finally in the towel. Now they’re stuck in this sclerotic state where updates to one platform aren’t automatically carried over to the others; new products, like risk management or FX tools, need to be built multiple times to accommodate different codebases; and transaction data that could inform risk scores and improve authorization rates circulate in siloes.

Incumbents responded to the mounting threat posed modern PSPs, not by streamlining their platforms to better compete, but by doubling down on megamergers. But buying scale never resolved the root problems that caused them to lose so much share in the first place. In a damning indictment of this acquisition-fueled strategy, FIS recently agreed to sell a 55% interest in Worldpay, valuing the latter at $18.5bn, a whopping 55% discount to the $43bn it paid just 4 years earlier.

But even compared to Stripe, arguably the most celebrated modern PSP of all, Adyen stands out. In 2022, Adyen reported 55% EBITDA margins on about $278mn of payment volume $483k of net revenue per employee (on today’s frontloaded employee base, which has grown by just over 50% from a year ago, Adyen is doing just $405k of net revenue per employee). By contrast, Stripe lost money on just $102mn of volume and $350k of net revenue per employee (they are supposedly on pace to hit $100mn of EBITDA this year…but including stock comp?)³. Some will object that the difference in profitability is due to compensation in the US being substantially higher than it is in Europe. That explains some, but not all of it. Even if Stripe’s expenses/employee were the same as Adyen’s last year, its EBITDA margins would have still been about 20 points lower.

For years the payments landscape was bifurcated between incumbents with meager growth who managed for margins and modern fintechs who reported losses but stole volume. Adyen avoided this trade-off. While legacy players acquired growth, adding to the jumble of platforms that made it so hard for them to innovate, Adyen grew organically, scaling one global platform. While modern fintechs splurged on pricey engineering talent, Adyen accelerated its hiring only in the last 3 quarters, when it was able to pick up for cheap talent that its peers were eagerly laying off. While PayPal groped its way toward a super-app and Stripe bet on sprawling initiatives to advance the grandiose vision of expanding “the GDP of the internet”, Adyen concentrated its engineering resources only on products its merchants wanted. In a sea of sloppy execution and distraction, the company exhibited a profound degree of focus, discipline, and frugality.

These unique cultural attributes powered eye-popping growth and industry-leading margins. Management announced last year it would be accelerating headcount to support growth, particularly in North America, an investment that would temporarily depress margins. But this was well understood and even celebrated (”look how long-term they are” “they zig when others zag” “capacity to suffer!”…that kind of thing). Then last quarter, Adyen cited “increasing competitive pressure in North America” as the culprit behind a dramatic deceleration in the region (net revenue growth in North America went from growing 36% in 2h22 to just 13% in 1h23) and everyone freaked out. It is one thing to invest ahead of growth. It is quite another to invest behind a structurally impaired market.

It soon became clear that price cutting at Braintree was most salient cause of the intensifying competitive environment.

Setting aside the execution and strategic issues that have plagued it, PayPal, who I once described as a “gangly fintech behemoth running on legacy tech whose long tail of fintech assets are united by vague super app ambitions but otherwise lack concrete product and ux cohesion”, is still a ubiquitous and trusted brand among consumers, who are supposedly more likely to convert into shoppers and spend more if the PayPal button is featured at checkout. In 2013 they acquired Braintree, an “unbranded” gateway that merchants could white label to accept payments across different merchant acquirers. A sideshow inside the PayPal complex, Braintree more or less operated as a separate entity for years. It was never prioritized for investment and lagged behind Adyen and Stripe, who have in the intervening years released a slew of additional products atop their payments baseload (from 2015 to LTM, Adyen’s processing volumes have grown from €32bn to €848bn while Braintree’s have grown from $50bn to an estimated ~$450bn). But after 8 years of trying, management claims that PayPal and Braintree now run on the same modern tech stack.

The technical integration is reflected in their commercial strategy too. Branded PayPal realizes much higher margins than Braintree, even controlling for differences in merchant mix, mostly because the former doesn’t pay interchange fees, the biggest “cost of goods” for a payments facilitator, on the 20% of PayPal transactions that bypass card rails. Moreover, according to Lisa Ellis from MoffettNathanson in her interview with Stratechery, a disproportionate amount of the other 80% flows through debit, whose interchange fees are much lower than credit. But the PayPal brand was also musty and saturated and facing competition from a host of alternative e-wallets. So to stoke branded adoption, PayPal began loss-leading with Braintree, giving the latter away at ~cost to merchants who also featured the PayPal checkout button. This maneuver has had a predictable effect on Braintree volumes, which swelled ~30% YTD to ~$450bn after growing 42% in 2022 (by comparison, in 2022 Adyen’s volumes grew 49% last year to $829bn while Stripe’s volumes grew 26% to $817bn).

Loss-leading with Braintree to boost branded PayPal adoption poses a conflict since PayPal is incentivized to push its own wallet while Braintree is compelled to integrate with all the most widely used payment methods. Braintree can choose to be agnostic about payment methods, but pricing at cost means that volumes processed through credit cards, Apple Pay, or any other medium except PayPal, are destructive to unit economics. Alternatively, Braintree can prioritize PayPal at checkout but doing so renders them less competitive with PSPs who feature a wider range of options.

Whether cross-subsidization proves viable ultimately boils down to how much shoppers prefer PayPal over other competing payment options. The incremental gains from using rock-bottom take rates at Braintree – take rates that don’t make sense absent a lift in branded PayPal volumes – to get the PayPal button prominently tattooed on more checkout pages could be more than offset by the incremental losses generated from transaction volumes shifting to Apple Pay or Shop Pay as online commerce continues to migrate to smartphones. It doesn’t matter if the PayPal button is featured in more places than it already is if people use it less. Even if branded volumes grow in aggregate, a big enough mix shift toward competing payment options intermediated through Braintree across the entire merchant base would still adversely impact profitability. So really what this strategy amounts to is doubling down on the PayPal wallet and, along with that, a bet on the associated investments and initiatives that the company has for years pursued to maintain branded PayPal’s relevance in eyes of consumers. Same as it ever was.

Will shareholders stick around for more of that? They’ve seen the stock crater by nearly 80% from its peak just more than 2 years ago. Transaction profits have declined from 1.3% to 0.89% of volumes since 2020 as margin dilutive unbranded volumes, fueled by aggressive pricing, have outpaced branded. Meanwhile, the company’s outsized stock based comp and flat margins seem out of sync with renewed concerns over profitability.

Braintree is rumored to have started cutting price ~1.5-2 years ago, which coincides with the marked deterioration in transaction margins and acceleration of Braintree volumes. That enterprises have become particularly cost-conscious and receptive to easy wins in the last few quarters could be why Adyen didn’t see much of an impact until this year. PayPal insists that unbranded margins will improve as Braintree caters to more smaller merchants (net revenue spreads are higher for low-volume customers), expands internationally (interchange costs are lower in Europe than in the US), and cross-sells ancillary FX and risk management services. In fact, they expect transaction margin dollars, which contracted in 2022 and ytd, to grow again in 4q. But that’s going to hard to pull off if they keep taking Braintree pricing down in the US, which still accounts for the vast majority of their volumes. So it’s possible that as low as Braintree pricing is today, it won’t be driven even lower from here, in which case the migration of processing volumes to Braintree should start to slow.

Still, you don’t necessarily want to bank on that? Even setting aside vigorous price cutting by Braintree, North America is an insanely competitive market and has been for decades. It is unrealistic to think that Adyen can deliver consistently better authorization rates here. Even with local acquiring and a modern tech stack, they are likely at a structural disadvantage to JP Morgan Chase, who issues about 20% of credit cards in the US and a similar share of merchant acquiring. If Adyen claims data advantages by owning the gateway and acquiring bank, how much better off still is JP Morgan, who owns the gateway, acquiring bank, and the card issuing bank, for the significant minority of US transactions intermediated through its cards? I would think that Braintree routing a Chase BIN to Chase produces systematically better auth rates than Ayden processing a Chase BIN internally? And then of course, Chase can and does loss lead with payments processing to win profitable banking and treasury management business.

But parsing performance by geography alone without considering broader strategic context is missing the forest for the trees. Adyen was carried to North America by its multinational merchant base and developing a presence there lent weight to its founding premise of offering local payments anywhere in the world on one platform. They have an opportunity to, in a sense, do some loss-leading of their own, only across geographies instead of payment layers, tweaking pricing in North America in service of a broader relationship that encompasses all the other regions, where payments is less commoditized and more fragmented.

The unique value that a PSP offers varies by use case and region: digital-only payments is a commodity in North America and somewhat less so in heterogeneous region like Europe, where payment types and compliance requirements vary by country. Marketplaces and platforms like eBay, Shopify, and Airbnb have complex global acquiring needs, but they are also tech companies at heart who have the resources and know-how to handle risk and KYC functions in-house, as well as the scale to negotiate favorable rates. Adyen is going after SMBs by plugging into the platforms that increasingly host them, thus converting SMB acquisition into a familiar enterprise sales motion. Its marketplace/platform product, Adyen for Platforms, has seen volumes explode from €3bn in 2019 to €102bn LTM, in large part due to eBay, a customer that Adyen stole from PayPal. Net revenue has got to be razor thin as a fraction of volumes but also very incrementally profitable on Adyen’s largely fixed cost base…plus, payments creates an opening to cross-sell lending, issuing, and other high margin ancillaries to the platform’s merchant base.

Where Adyen stands apart from the pack is in its ability to serve sophisticated global retailers who are not tech companies themselves but are tech-forward enough to embrace omnichannel. It is far easier for large merchants to do BOPIS, online sales with offline returns, and otherwise create 360 customer profiles by consolidating on a single platform that traverses both digital and physical commerce. Volumes from Unified Commerce (omnichannel) have exploded from just €45bn in 2019 (19% of total processed volumes) to €225bn (27%), with the point-of-sale component ballooning from €17bn to €124bn. Over the last year, UC and POS volumes have grown 47% and 54%, respectively, outpacing the 30% growth rates posted by Digital-only, where the competitive pressures from North America have been most acutely felt. Besides the deluge of payment volumes it delivers, omnichannel is a hook for more enduring and expansive engagements. Offline commerce is stickier – it is much harder to swap out physical terminals than it is to direct online volumes to a competing PSP – so a retailer who uses Adyen for point-of-sale and takes omnichannel seriously will also be more inclined to route online volumes to Adyen as well, assuming no systemic deficiencies in authorization rates. As customers consolidate more of their flows, Adyen can monetize other hooks like risk management, card issuing, capital, and payouts.

So if North America is a race to zero, then that raises the question of who is providing the most value elsewhere. Everyone will do FX services, risk alerting, issuing, and the rest. But Adyen seems uniquely placed to claim end-to-end local acquiring across many regions at once, online and offline. To put it another way, Adyen is more an “N of 1” in the thing they are known for than other PSPs are in the things they are known for. They are also more unique and moated relative to other PSPs than PayPal is to other wallets. Braintree is predominantly focused on the US and doesn’t have full control of the merchant-side payments stack. Mollie is geared to SMBs in Europe. Stripe got started with developers and SMBs and has been trying to climb upmarket, but isn’t yet on par with Adyen in addressing global enterprise use cases and lags behind Adyen in omnichannel. Checkout.com is a European rival that also targets large enterprises with a modern tech stack and local acquiring licenses. While often overlooked in payments conversations – I think most investors would be surprised to know that Checkout was valued at $40bn at its Jan. 2022 peak before it crashed with everything else in fintech – they actually seem like a promising contender to Adyen, at least when it comes to digital-only flows. Meanwhile, I can think of more reasons why Apple Pay and Buy with Prime are better positioned to take share from PayPal than vice versa.

In a nutshell, Adyen’s advantage exists at the intersection of several capabilities. All payments are local but the large enterprises and platforms that Adyen serves are global. There are lots of PSPs out there that can do local processing, but they don’t have Adyen’s global scale. There are incumbents with global scale, but they don’t run a unified platform through a single integration. Payment processing in North America may be a commodity, but local acquiring scaled globally is not. If Braintree insists on pricing at cost, then Adyen will have no choice but to get off its horse and take prices down too. But in my opinion they are better placed to sustainably do so than PayPal both because their cost structure is leaner and because local acquiring at global scale through one platform is a far more differentiated prop that earns them the right to take profits elsewhere.

That Braintree price cutting has had such a sudden and material impact on Adyen volumes in North America has also revived concerns that payments provision is being further commoditized by orchestration layers that make it easy to route transactions to different PSPs based on authorization rates and cost. Disintermediation and aggregation have been ongoing themes since the dawn of electronic payments. Card networks like Visa and Mastercard replace sequestered credit lines at each merchant with one facility that works across all of them. E-wallets like PayPal encompass ACH, account balances, or credit cards. PSPs like Stripe or Adyen subsume wallets, credit card, and bank transfers, payment methods that in turn can be intermediated by an aggregator like PPRO in foreign markets. Gateways provide choice across different acquirers, orchestration platforms do the same across different gateways and PSPs.

The pitch from orchestrators is that responsible merchants should avoid locking themselves in to any one PSP or gateway. The subtext, of course, is that in doing so, the merchant should be more dependent on them. I’m skeptical that PSP orchestration is solving a hard enough problem to warrant a seat at an already crowded table. The sophisticated retailers and platforms that Adyen serves, the type of customers that most value orchestration, already process through dozens of PSPs and will have built routing logic in-house. They might engage a primary acquirer to handle like 60% of flows in a certain region while spreading the remainder across several others, dynamically re-directing volumes if authorization rates or processing costs trip certain thresholds, routing flows based on a PSP’s track record of handling various transaction types.

None of this new. Orchestrators have been around for decades. Enterprise merchants have always plugged into multiple PSPs. And yet, with 80% of payment volumes coming from existing customers and less than 1% of volume churning every year, a significant chunk of Adyen’s growth has come from consistently taking wallet share⁴. Those share gains have coincided with Adyen acquiring banking licenses and offering omnichannel capabilities in new territories, as what might begin as a limited online-only engagement with a retailer in one country eventually expands into an omnichannel relationship across several. Much as Google has retained search share despite competition being just “a click away”, Adyen has consumed a greater share of volumes from merchants who could rather easily re-route to any number of PSPs they connect to, validating the package of competitive auth rates and ease of integration and settlement reporting that Adyen differentiates on.

Given that enterprise customers continue to engage with myriad PSPs, I suspect there is more wallet space for Adyen to carry its advantages into than there are existing volumes for orchestrators and alternative PSPs to route away, particularly since the value of a single integration that handles and formats settlement data in a uniform manner across channels and countries grows as more volumes are concentrated on it. That doesn’t mean a global merchant will sole-source on Adyen (this will never happen). But I can entertain a scenario where an enterprise merchant or platform defaults to the PSP that can solve the hard problem while a jump ball on the commoditized parts, like North American digital-only commerce. In which case, we could see orchestrators starved of oxygen as Adyen, who I’m pretty sure doesn’t plug into third party orchestrators, continues to take wallet share.

I bought some shares of Adyen last month. It’s a small position. That I even own it at all basically reflects my view that Adyen has a uniquely strong value prop for a large slice of global commerce and is run by a thoughtful founder with skin in the game⁵ who will figure out how to convert that position into shareholder value. Payment processing isn’t a franchise like Moody’s. It’s a fast-moving, competitive space that a lean and agile organization like Adyen will maneuver through better than its heavier competitors.

That I haven’t sized it bigger reflects my uncertainty about where profitability ultimately settles and how large the addressable market is for the particular set of capabilities Adyen brings to bear. Management stands behind its long-term 65% EBITDA margin guidance and says it could get there “very, very quickly” if it weren’t investing to support growth. Whether that’s true depends of course on the competitive posture in North America, how Adyen responds to it, and to what degree take rate compression spills over into Europe. Maybe Adyen remains firm on price, PayPal reverses Braintree loss-leading to right its transaction margins, and North American pricing normalizes? Maybe Adyen relents on price and parlays US volume wins into more encompassing global engagements, such that lower transaction margins are offset by volumes to produce similar amount of profits? Moreover, the part of the global payments TAM that Adyen can attack at durably compelling margins is constrained by the fact that its right to earn is grounded in the unique way it serves global merchants with multifaceted flows who think of payments as something more than a commodity. Like, it would be wrong to compare Adyen’s $145bn of point-of-sale volumes against tens of trillions of global offline volumes since only an unknown subset of the latter lands in Adyen’s sweet spot.

All that said, I don’t think you need 65% margins and heroic growth rates for the stock to work from here. Over the next 5 years, if we assume growth is 0 in North America (vs. 33% y/y growth LTM), 15% in EMEA (20%), 20% in Latam (24%), and 25% in APAC (37%), I blend to about 14% on a consolidated basis (compared to 25% LTM and 33% in 2022). This gets us to about $2tn of processed volumes, about where Chase, Worldpay, and First Data are today. Assuming 70% of incremental revenue drops to EBITDA, I land at 58% margins (vs. 63% in 2021). With 80% of EBITDA dropping to net earnings, I can pencil 12%-13% returns at a 25x terminal multiple, including accumulated free cash. Now, even after losing 60% of its value, Adyen still trades at 36x (hopefully depressed) trailing earnings – a huge premium to legacy peers, who trade at low-teens, and in the same ballpark as consensus compounders Visa and Mastercard – so it’s not like the market thinks the company is another payments shitco. I don’t think Adyen is a screaming bargain here. But it’s interesting enough (for me) to get started.

Disclosure: At the time this report was posted, accounts managed by Compound Insight LLC owned shares of Adyen and Moody’s. This may have changed at any time since.

1. BINs and Interbank Card Association (ICA) numbers are unique numbers assigned by Visa and Mastercard, respectively, to member banks that allow those banks to clear and settle transactions
2. in 2019, management disclosed that 70% of its volumes are processed end-to-end
3. net revenue estimate from The Information
4. that their merchants are growing contributes too, of course.
5. CEO Pieter van der Does owns about 3% of the company, ~$700mn of stock.