Marqeta, Toast and the Enterprise segment
Enterprise must be balanced with higher-margin segments
Key insights in the post
The Enterprise Segment has almost opposite characteristics from small business
Low revenue per transaction but high volume per client
Low distribution costs and low churn
Low propensity to buy bundles of ancillary products
Typically, a processor needs Enterprise volume for scale but Small Business volume for profits. Toast & Marqeta are approaching Enterprise from different starting points
Toast
Toast started with Small Business
It has leading share in the US restaurant vertical
It sells a high-margin ISV bundle of Software + Payments + VAS
It is now going global and diversifying into other SB verticals
Toast is entering Enterprise with a largely Software-only proposition
Enterprise payments revenues are low anyway
Enterprise accelerates software revenue growth without adding much to R&D costs
Marqeta focused on Hypergrowth Fintechs who grew to become Enterprise
Tiered pricing gives automatic volume discounts, squeezing margins
The biggest clients graduate to “Powered by Marqeta” (PxM) where revenue per transaction is low. PxM clients may account for ~80% of revenue
Cash App alone accounts for almost half of revenue and more than half of volume, despite deep discounts and insourcing many services
Some smaller clients are also growing fast, but that moves them up the tiers to lower marginal prices and closer to the PxM threshold
Conclusions
Enterprise can be lucrative under some circumstances
When it is an adjunct to a high-margin SB business
When services are differentiated enough to maintain margins
Toast has been successful with the former model, and Marqeta tried for the second model but could not maintain margins
It is not clear how Marqeta can adjust given its customer base of hyper-growth Fintechs. It needs less impressive clients!
Introduction
I frequently comment on the chart below to demonstrate the implications of serving different B2B segments:
The ideal situation is to have lots of Small Business volume to earn profits while building scale, and then move upmarket. While distributing to small business is expensive, the margins make up for it. To serve Enterprise, you need massive scale because the revenue per transaction is so low – however the distribution costs are relatively modest.
I think Marqeta, an innovative company I have long admired, has suffered financially because of the segments it serves rather than its business model per se. In contrast, Toast (another favorite) built scale in Small Business and only then moved upmarket. Their results show a more positive trajectory. I thought this comparison would make for an interesting post, the lessons from which can be applied to many other Fintechs.
Enterprise is hard
I describe the Fintech model in Small Business as Software + Payments + VAS, leading to high ARPU. That formula applies only to the small business segment. Once a merchant grows above a certain threshold the three-prong approach is harder:
Payments: Price per transaction drops as clients get larger. This most evident in Acquiring where pricing shifts from the MDR model (bps of spend) to the IC+ model (pennies per transaction)
VAS: Use may drop to zero, especially for any lending products, as larger companies can borrow more cheaply from banks or capital markets
Software: Still needed by large clients – but pricing drops with size
To afford Enterprise customers, a Fintech needs that segment’s software revenue to be incremental to a full-service Small Business base. The SB segment foots the bulk of the bill while Enterprise generates marginal revenue on the same code base.
Consider Fiserv. They helpfully disclose the difference in revenue between their Enterprise segment and their Small Business segment:
Although Fiserv does not disclose the TPV for each segment, an SB transaction could generate 50-100¢ in payments revenue, plus value-added services like Clover Capital. In contrast, an Enterprise transaction may generate under 5¢ include VAS like fraud detection. I would be willing to bet that the Enterprise segment processes at least 3x the volume yet SB generates 3x the total revenue. Here is a post on Fiserv’s efforts to expand in SB via its Clover solution.
14% transaction growth in enterprise only generated 8% revenue growth – which may be because most of the growth came from the biggest merchants who pay the lowest fees. In contrast, SB revenue growth and volume growth are roughly the same – because SBs buy the full bundle on Clover.
The flip side of this is distribution & servicing costs. SB distribution is expensive, not only because SBs are so dispersed but because the segment has significant churn. Enterprise may be expensive to capture (usually via a formal RFP process), but they are the gift that keeps on giving, as defection is rare — although price reductions are common on renewal.
Another interesting comparison is Adyen vs. Stripe.
Adyen started in the MNC segment serving cross-border digital goods companies that needed wide geographic and payment method support (e.g., online gaming and legal gambling, streaming media, etc.). Effectively, Adyen started in Enterprise and only recently began moving down-market by serving platforms. They could survive at Enterprise price points because they got a premium for their global coverage and payment method coverage
Stripe started by supporting US eCommerce startups. That was a lucrative segment as they got an SB premium and an eCommerce premium. Stripe moved upmarket as their hyper-growth clients got bigger and globalized – needing support more like what Adyen provided. In other words, Stripe supported their legacy clients as those clients became Enterprise. Not too much later they started winning Enterprise business directly
As Adyen moves down-market their margins will go up but, so does their distribution costs while as Stripe moves up market its margins will compress but its volumes will grow in bigger chunks. Both companies will still get a premium for processing eCommerce transactions, global footprint, and comprehensive local payments methods – and for their superior platforms built on modern tech architectures.
These examples demonstrate that moving upmarket fundamentally shifts the type of revenue earned and the associated margins. We will see how this plays out with Toast and Marqeta.
Toast is moving upmarket after winning a large SB base
Toast has ~148K locations, the vast bulk of whom are single-location US restaurants. They supplemented core payments and software revenue with Value Added Services (VAS), such as Toast Capital. After winning in that arena they recently started pursuing global expansion, vertical expansion, and upmarket expansion. Here are some of my prior posts that discuss this in detail: Global expansion, competitive position, share in restaurant vertical
Toast understands that large enterprises generally won’t take bundled payment services – and, if they do, it is at much lower price points; but, having covered fixed costs from SB, capturing software revenue from Enterprises represents incremental margin.
“A Location can use Toast payment services, which we refer to as a Toast Processing Location, or for select enterprise customers, not use Toast’s payment services, which we refer to as a Non-Toast Processing Location.
Here is a comparison of key metrics for the first six months of 2024 & 2025
Subscription Services Revenue is outgrowing Financial Technology revenue substantially, likely because Enterprise clients only use Subscription Services.
Toast gives very little insight into the location mix, i.e., how many are “Non-Toast Processing”, how many are not restaurants and how many are not US. But there are some early signs of an Enterprise impact in the ARPU growth rates for Subscription Service Revenue and Financial Technology Revenue:
Financial Technology ARPU is growing slower than inflation on both metrics whereas Subscription Services ARPU is growing faster.
Financial Technology results could be due to disproportionate down-market winds in addition to Enterprise wins. Smaller merchants produce less payments revenue than standard merchants, although revenue share of TPV should be similar
Subscription Services results can only be from more “Non-Toast Processing” merchants who only take Subscription Services. More small merchants should drive this metric down – particularly on the “Growth per location” metric, so it is more likely from more Enterprise merchants
We have evidence that the Enterprise Subscription Services revenue is high margin since Cost of Revenue (CoR) growth was much slower for Subscription Services:
Financial Technology CoR rose slightly slower than revenue (23% vs. 24%)
Subscription Services CoR rose much slower than revenue (26% vs. 37%)
Further, R&D is the main operating cost associated with Subscription Services and that rose only 4% YoY.
We don’t know how may Enterprise merchants Toast has, but they did disclose:
“… enterprise, international, and food and beverage retail passing 10,000 live locations”.
They also announced one specific chain signing up with 1,300 locations.
Each location generates roughly $6K in annual Subscription Services Revenue. The 1,300 location chain would be worth $7.8M in revenue at that rate. Volume discounts may drive that number lower but complex needs may drive it higher. Either way, most of that revenue is incremental and it likely comes with lower CAC per location acquired. According to the Census Bureau Establishments survey there are 1,200+ chains in the largest category, although only a fraction of these are at 1000+ scale.
The overall lesson here is that serving Enterprise requires a balance with serving smaller merchants. Enterprise is attractive as an incremental segment, but not so much as a primary target.
Marqeta was dragged upmarket by its hyper-growth clients
Marqeta has two service models, one for its bigger customers and one for its smaller ones. This is described in their 10-Q, but I edited and formatted the text for readability:
In general, the MxM model earns a share of interchange while the PxM model is paid via transaction fees.
Marqeta has a “high class” problem: Most of its volume is from hyper-growth Fintechs in the PxM model. This is a challenge due to revenue sharing arrangements [bolding added to 10-Q text]:
“Revenue Share is generally computed as a percentage of the Interchange Fees earned on processing volume and is paid to our MxM customers monthly. Revenue Share payments are recorded as a reduction to net revenue. Generally, as customers' processing volumes increase, the rates at which we share revenue increase.”
As clients scale, Gross Margin declines. The faster those clients grow, the faster margins erode. Eventually they may transition to PxM where per transaction revenue is a step function lower.
A reciprocal phenomenon happens on the cost side, but at much lower magnitudes. For example, BIN-sponsor fees from Sutton Bank have similar tiering — that benefit shows up within “Cost of Revenue”; but, Sutton only reported ~$15M in first half income from this source. That is under 20% of Marqeta’s Cost of Revenue.
The trouble started with Cash App, which in recent history accounted for almost 80% of Marqeta TPV but has the tightest margins of all. Here is a prior post on Cash App growth.
Marqeta even changed its accounting treatment in 2023 to replace the Gross Revenue/Cost of Revenue treatment for PxM clients (mostly Block at the time); now they just report the net. In Q3 23, the QoQ adjustment reduced Net Revenue by 53% but Gross Profit only declined by 14% and net losses actually declined! Opex actually declined as much as Gross Profit, so the whole contract was arguably earnings neutral. In other words, 80% of Marqeta processing was at best break even at the time.
Assuming Block accounted for 80% of TPV, the revenue lost was only earning 3bps Gross Profit on TPV. The rest of the book was likely earning 20bp+. In their Q2 25 10-Q Marqeta outlines today’s Revenue concentration [bolding added]:
“… Total platform services, net revenue increased by $41.8 million, or 18%, for the six months ended June 30, 2025, compared to the same period in 2024. The overall increase in platform services revenue was primarily driven by a 28% increase in TPV, partially offset by unfavorable shifts in our card program mix, particularly the expansion of programs where we provide processing services with minimal or no program management …
… TPV for our top five customers … grew by 19% for the six months ended June 30, 2025, compared to the same period in 2024. TPV from all other customers, as a group, increased by 67% in the six months ended June 30, 2025, compared to the same period in 2024 …”
I calculated the mix based on this data. The shaded boxes are calculated:
For the first half, the top 5 likely accounted for 81% of TPV. The biggest of these is Cash App which alone used to account for 75%+ of revenue and TPV. For years, Cash App was outgrowing the rest of the customer base, accounting for ever greater shares of volume. This was so pronounced that I used to quip that Marqeta’s losses represented a wealth transfer from Marqeta shareholders to Block shareholders. Today Cash App is just under half of Revenue.
This section of the 10-Q had some odd discrepancies:
For TPV they show a Top 5 (see above) but here they limit disclosure to a Top 3. #4 & #5 together may account for another 5%+ if we compare the H1 total here (76%) to the TPV table above (81%)
Cash App share was for “Net Revenue” while #2 & #3 were for “Accounts Receivable Balance” a term otherwise absent in these sections of the 10-Q. I assumed it is used here as a synonym for Net Revenue, but I could be wrong
Cash App growth is YoY but #2 & #3 is for Q2 25 vs. Q4 24 (i.e., six months), which means their YoY growth may be twice as high. #2 in particular may be outgrowing the “All Other” category in the TPV table and accounts for the same share.
The take-away is that despite the modest decline in Cash App share, revenue concentration is increasing. Their top clients are Silicon Valley royalty like Klarna, Affirm, Cash App, Uber & Door Dash. Marqeta’s problem is that these big clients grow too fast!
This works financially if there is an inflow of small, slower-growing customers with fat margins. We can see in the TPV table that volume outside the top 5 grew 67% YoY. These “All other” customers will graduate into higher revenue sharing tiers and some may eventually flip to PxM. There just aren’t enough small clients entering the customer base at the bottom to replace the Net Revenue erosion from the bigger, hyper-growth companies.
Conclusions
This analysis demonstrates that serving Enterprise is a mixed blessing that only works in narrow circumstances:
When the provider is differentiated enough to maintain pricing premiums, e.g., Adyen for cross-border
When Enterprise represents incremental revenue above a solid Small Business or Middle Market customer base (e.g., Toast)
Unfortunately for Marqeta, their best MxM clients grew to Enterprise scale before Marqeta was ready to support the PxM model financially. So, margins compressed on the bulk of their book, offset by a modest long tail with fatter margins. Toast still supports a big small business customer base, treating Enterprise as incremental.
In Marqeta’s case, TPV growth is giving off a false signal. If TPV growth derives from a Fintech’s biggest clients, it makes margin compression worse; if it comes from long-tail clients, it still compresses margins due to tiering. Only if growth comes from new, smaller clients does it provide the kind of high-margin growth needed – but only if they stay small.
It hard to see what Marqeta can do about this. They can’t raise prices, they can’t tell clients to stop growing, and they can’t tell scale clients to stop insourcing select services. They need more boring clients, but can they even find enough small Fintechs with static volumes to balance the book?