Klarna isn't a credit card killer

Pay-in-4 is a great product for credit-constrained consumers only

Key insights in this post

• Klarna recently withdrew its IPO filing given market turmoil

• The S-1 was rife with anti-incumbent rhetoric that contrasted the benefits of Klarna’s Pay-in-4 model with the downsides of Credit Cards, and banks in general

• The polemics are unnecessary as Klarna is a glowing example of financial inclusion: It provides a little bit of credit, to credit challenged consumers, over short periods, for free. And it gets merchants to fund this voluntarily!

• Many of the comparisons to US credit cards are not fair

  • For consumers Klarna is only a peripheral competitor to credit cards

    • They serve credit-constrained consumers, either debit-centric or low-line

    • They have no appeal to Transactors who make up half the cardholder base

    • They don’t offer enough borrowing capacity for creditworthy revolvers, and where they do, it is at credit card-like APRs

  • For merchants, Klarna are not generally cheaper than cards

    • They are often as expensive as cards for equivalent consumers

    • They are promoting virtual card purchases that are more expensive than all but the highest-end credit cards

• Their business model is unlikely to challenge Credit Cards going forward

  • Their underwriting strategy does produce low losses, but the product structure accounts for most of that, not superior underwriting

  • Their data strategy has more credibility than most, but lacks scale

  • They will find it challenging to raise low-cost US deposits, as they plan

  • They face energized competition, both from other Pay-in-4 providers and alternative borrowing options

• Given the success of their core product, the polemics just aren’t necessary

Introduction

I have been avoiding this post since Klarna published its S-1. I am on record as saying that BNPL is a lending product not a payments product, so I thought I should stay away. Also, I genuinely think Klarna is an admirable example of financial inclusion – they lend small amounts of money to credit constrained consumers for short terms at no interest — and get merchants to willingly pay for it. Magic. Another excuse was that the initial S-1 didn’t have 2024 financials. It would be hard to evaluate much without recent numbers. The S-1 left space for them, but those spaces were blank.

But I went back and reread the S-1 after Klarna withdrew it. It turns out Klarna is a payment product 23% of the time, accounting for 16% of GMV; That is how often users “Pay in Full” with Klarna products. So, just like other payment methods, at least some customers use it just for transacting. Those 2024 blanks are now filled in, so I no longer have that excuse either. Finally, since they have withdrawn the filing, my comments can be viewed as constructive criticism – which they are intended to be.

As usual, I will restrict my analysis to the US as I don’t have enough insight into their European core business. My overall theme is that their constant comparison of themselves to Credit Cards is inapt:

• In the US, Klarna is only a peripheral competitor to credit cards today

• Their strategy won’t close the gap materially

• They face energized competition, both from other Pay-in-Four providers and alternative borrowing options

In other words, their anti-incumbent rhetoric is off-base and distracts from what are admirable accomplishments.

Klarna is only a peripheral competitor to US credit cards

The CEO’s introduction to the S-1 is a polemic against the credit card industry and banks in general:

“… Yet, somewhere along the way, traditional banks traded this principle for profit, losing sight of what truly matters. Instead came financial engineering, raking in profits through late fees, overdraft penalties, revolving debt traps, and countless other tricks designed to exploit their customers. Trust in banks has never been lower.”

“… we took the opposite approach. Late fees? We minimized them. Revolving credit? Eliminated. Our Buy Now, Pay Later (BNPL) feature has saved consumers billions in interest compared to using credit cards.”

“Klarna, by contrast, lets you decide on a purchase-by-purchase basis whether to pay now with debit or use credit with clear, fixed terms. It's a difference that keeps the average Klarna balance at $87, compared to the $6,730 average credit card balance in the United States …”

“Klarna … is the choice of a new generation-one smart enough to avoid credit cards and banks that rely on outdated tricks.”

So much to unpack! Much of my critique comes down to a key fallacy in assumptions: BNPL users in the US are typically credit constrained: they either don’t have credit cards at all or they don’t have enough line on the cards they have. BNPL is their only credit option. They may be younger (“a new generation”), but not always.

On page 25, Klarna lists global GMV at $105B for 2024. They do not provide a geographic GMV breakdown, but, the US accounted for 30% of revenue. If revenue tracks spend, Klarna would have $32B in US GMV. That is out of 2024 US credit card spend of $6.1T – or about 52bp. Card purchase volume grew 6%, generating 9 Klarna’s worth of new volume.

More than half of Klarna users are debit-centric. US 2024 Debit spending was $4.6T, so Klarna would account for 89bp basis points of Debit spend. But that is double-counting because most US Klarna transactions trigger 4 US debit transactions. Rather than displacing them, it quadruples them.

Against Combined Debit+Credit volume Klarna is ~30bp.

In contrast, my old employer JP Morgan, had $1.3T in credit card spend and grew that at12% -- Klarna was at 14% growth globally. In other words, JPM alone added 4.5 Klarna’s worth of credit card volume in 2024 – plus debit card growth. From a consumer choice perspective, US credit card holders are not convinced.

I don’t mean to diminish Klarna’s accomplishment, which is impressive, but it is not even a little comparable. Let’s look at how Pay-in-four compares to credit cards from both the consumer and merchant perspectives:

Consumers

In my consulting past, I analyzed primary DDA account data to identify what kinds of DDA customers were using BNPL. A DDA is Primary if the consumer gets Direct Deposit and pays bill from that account, so we could see which accounts were paying a monthly credit card bill and which weren’t. And we could see from debit transactions which ones were paying BNPL loans.

At that time, 90%+ of BNPL users didn’t have credit cards at all. The remainder had cards, but carried a balance and were clearly using BNPL to avoid adding to that balance. The mix has shifted since then and may be closer to 50/50 between Credit Card holders with constrained credit and debit-centric consumers with no Credit Card. But there was simply no evidence that a “new-generation” was avoiding credit-cards – it was the card issuers that were avoiding them! At least, until their credit scores improve. See this post from just a few weeks ago for a deeper dive.

More affluent cardholders don’t revolve on their cards, so get no benefit from BNPL. The industry calls these cardholders “Transactors” and they account for more than half the card base.

• They get 30 days free float, on the full purchase price – which is more than Klarna gives on Pay-in-four. In fact, Pay-in-four customers have paid back three-fourths of their loan before a Transactor even gets a statement. At best, BNPL gets a transactor an extra two weeks on 25% of the purchase price

• They don’t pay fees – except for super-premium cards where they get value for their fees, like lounge access

• They get rewards – the issuers actually pay Transactors to use their cards. Klarna doesn’t do that

• They get higher spend capacity to fund bigger purchases like furniture, travel, etc. These consumers generally pay-in-full each month, so underwriting each purchase is irrelevant. When Affirm was offering 0% APR installment loans for Pelotons and the like, Transactors did take advantage of it, but out of financial savvy, not necessity

Affluent consumers of all ages flock to these card products. They have almost universal penetration in this demographic. It is not clear to me why any Transactor would choose Pay-in-4 over one of these cards for any transaction.

“Revolvers” are the other half of cardholders. These cardholders carry a balance. They do pay high APRs (depending on credit quality) and some of them pay late fees and other nuisance fees. About 20% of cardholders overall paid a credit card late fee according to a Nerdwallet survey. The consumer can use Klarna to fund a larger purchase over a longer term but, those installment loans carry credit card-like interest rates and account for only 1% of transactions (5% of GMV).

Of course, the average revolving balance on a credit Card is higher than on Klarna because the issuers grant more credit to higher FICO consumers. Since those consumers already have higher creditworthiness than Klarna pay-in-4 customers, they are entitled to more credit. Would Klarna even lend $6K to a pay-in-4 customer? It’s current average loan is $87. At a credit limit like that a typical revolver would need dozens of Pay-in-4 loans.

Merchants

“We operate a sustainable business model defined by lower fees for both merchants and consumers relative to legacy payment networks, such as credit cards”

The bolding in mine.

For smaller merchants, this is a true statement. Klarna merchant fees are around 2.5%, although that likely varies by merchant size. Small business Acquirers charge somewhat more: Shopify, for example, charge 2.4-2.9%+30¢ for online card acceptance. Square charges 2.9%+30¢. So Klarna is a bit less expense — about 70bps less.

But larger merchants pay IC+ pricing, which means interchange is passed through and the acquirer charges a very low per transaction fee – less than a penny for the largest merchants. Even network fees are discounted. The lowest categories of Credit Card CNP interchange would cost 1.89-2.05% + 10¢. Adding network fees and acquiring fees brings the total cost on a $100 transaction to $2.16-2.22 compared to Klarna at $2.50. Klarna costs 18-22bp more. Of course, larger merchant likely negotiate lower Klarna MDRs as well. My guess is the costs are roughly comparable after discounts.

What Klarna is likely comparing to is premium card interchange, which is paid on Transactor cards and can reach $2.88 per $100 transaction with network fees and acquiring. But this is not apples-to-apples because those customers have a much deeper spend wallet – and Klarna brings few of those to the table.

Of course, over half of Klarna clients are debit centric, so they should be comparing to debit acceptance costs. Interchange on Regulated debit is 22¢+5bp. Adding network fees and acquirer fees might bring this to ~44¢ a transaction at $100 and ~34¢ for the typical $50 debit transaction. Klarna would charge $2.50 & $1.25 respectively. Exempt rates are higher but still well below Klarna rates. Exempt cards account for only 30% of volume.

In other words, for similar, credit-constrained clients at large merchants, Klarna is equivalent to credit cards and a lot more expensive than debit cards. In fact, Klarna economics wouldn’t work if debit was more expensive because the four debit installments they collect would cost more than the 2.5% they get from the merchant.

Finally, it is a bit ironic that Klarna is criticizing the cost of accepting credit cards when its own products use high-interchange cards:

“By integrating Klarna [Card] with Apple Pay and Google Pay, our consumers can use Klarna’s payment solutions wherever Apple Pay or Google Pay is available online in the United States …”

If you pay using the Klarna Card or Apple Pay, Klarna is using a network-branded virtual card which carries interchange almost as high as affluent cards. And unlike the normal Klarna fee, virtual card interchange is not negotiable by the merchant. They are just stuck paying it. These kinds of transactions are under 10% of GMV today but growing 50% faster than the core transactions.

Polemics aside, Merchants think Klarna lifts basket size, conversion, and repeat visits, and that is worth a premium to them. Claiming that it is somehow about cost when it isn’t actually hides the key message – Merchants value Klarna enough to voluntarily pay more!

The business strategy won’t close the gaps

Klarna makes claims about its business strategy relative to incumbents. Some of these comparisons are valid and some are overstated.

Underwriting losses

“Our underwriting process results in credit losses that are generally lower than the industry average: for example, our credit losses represented 0.47% of GMV in 2024”

“We also provide a small spending capacity that gradually increases as consumers responsibly spend more with Klarna, and clear and transparent repayment terms that encourage borrowers to repay on time. All of this distinguishes our financing solutions from market alternatives.”

“In 2024, our average balance per active Klarna consumer was $87 (compared to an average balance per credit card of approximately $6,730 in the United States in 2024, according to Experian), and average loan duration was approximately 40 days”

My understanding is that credit losses are fundamentally different for Pay-in-4 providers where application fraud is a bigger challenge that inability to repay. But as Klarna points out, their pay-in-4 loans last for 40 days, while credit card loans have indeterminate duration, but usually less than 24 months. Klarna is rightly proud of this accomplishment. But Klarna is only really lending 75% of the balance since the first 25% payment occurs immediately. Therefore, the 47bps is more like 63bps for the amount actually at risk.

Another difference is bi-weekly repayment. Three-fourths of the balance is paid back by the 30-day mark – before a credit card statement is even issued. Klarna are only on the hook for the final 25% of the loan after that first 30 days. Reducing risk further is that all installments are on autopay via debit card. This model eliminates much of the repayment risk because the money is paid back quickly and automatically. Perhaps there are lessons in this for sub-prime card issuers.

Of course all this means that the consumer needs to apply again for subsequent purchases. This is not as convenient as a card. A consistent criticism from activists and regulators is that repeated BNPL use is not that dissimilar from other kinds of debt traps. I don’t think this comparison is fair as Klarna is not charging interest, so where is the “trap”? But, the criticism is out there and could lead to regulation at some point.

Impact on banks

The overall tone of the S-1 is that Klarna intends to displace legacy incumbents, but the mechanics of the Pay-in-4 product actually help most of those banks.

As we know, the Durbin amendment limits the interchange a large bank can earn on debit. Large, “Regulated” banks generate ~70% of debit volume but their interchange is capped at 22¢ + 5bp per transaction. The way Pay-in-4 works, the customer makes four debit payments on each purchase – each for one-fourth the purchase price. So, a $100 purchase results in 4 x $25 debit transactions or 93¢ in interchange instead of 27¢. This is a windfall for those large banks, a lift of 66¢.

Interestingly, it helps small banks much less. The “Basic” Visa exempt interchange rate for eCommerce is 160bps + 15¢. Because the bps are the same whether the transaction is processed in a single payment or 4 payments, the lift is smaller. Exempt banks would get $1.80 on that $100 payment without Klarna and $2.25 with Klarna — a lift of only 45¢.

So Klarna is a DDA bank’s best friend.

In the future, I expect Klarna will move from Debit to ACH as a collection method as it is much less costly, but also more risky. Open Banking will reduce the risk, but it adds cost, so the net savings are hard to calculate. That will take that interchange windfall down to zero.

Klarna does hurt the revenues of credit cards issuers when their cardholders use Pay-in-4 to avoid adding to their card balance. Pay-in-4 charges no interest where the issuer is likely charging north of 20%. But a relative handful of issuers focus on this segment, so the pain is concentrated.

Data Strategy

As my longer-term readers know, I don’t think payments data is all that valuable for marketing purchases – but, Klarna may be an exception. Like Affirm, AfterPay/Block and others, Klarna gets SKU data on the specific items they finance. They use SKU for fraud and underwriting assessments. Credit Card issuers only get the total purchase amount and merchant category. Klarna gets one SKU per transaction, not multiple SKUs per basket.

“… in 2019, we started to meaningfully scale our advertising solutions, which personalize the commerce experience for our consumers by using our vast proprietary data set, including data they entrust to us”

“We have also scaled our advertising revenue from approximately $13 million in 2020 to $180 million in 2024.”

That is a 90%+ CAGR. But how much is their “vast proprietary data set” involved? Here is how Klarna describes their advertising efforts:

“We offer brand, search and affiliate solutions to advertisers such that they can connect with consumers across the commerce journey. We also allow merchants to reach consumers in a commerce-centric environment, which we believe is the most effective place to reach consumers.”

“Search and Affiliate” are the key words here. From Investopedia: “Affiliate marketers get paid a commission for referring customers to companies where they make purchases.” Many newsletters and blogs earn affiliate fees from Amazon, simply by putting links on their sites. So does my wife’s nonprofit from their website.

In Klarna’s case, loyal customers may start a shopping journey in the app and search for a specific good available at a merchant that accepts Klarna; They are then transferred to that merchant. No customer data is used in that transaction. The customer initiates the search and Klarna executes the results. Affiliate fees are how Klarna get paid for that search traffic.

Brand advertising fees may be targeted based on customer data, but the biggest likely revenue source are affiliate fees. Klarna does not reveal what share of these revenues are the Brand kind versus the Search kind.

Of course, data is only valuable if you have enough of it. We saw in my PayPal engagement analysis that their wallet wasn’t used often enough to build a truly predictive purchase database. The Payments Data Illusion discussed how Cardlytics, with massive amounts of debit purchase data, had not built a particularly profitable business from it (market cap is currently $75M).

All the banks that contract with Cardlytics earn advertising revenue as well. If I recall, Cardlytics distributes about as much income to banks as it keeps as its own revenue, so the banks as a whole got almost $300M in revenue in proportion to their debit portfolio size. The Cardlytics network overall has a MAU of 167M, mostly from debit centric consumers — the same ones that use Klarna.

So even with relatively passive techniques, any company with transaction traffic can earn at least some of this kind of revenue. But does Klarna’s access to SKU data allow it to build predictive analytics that earn a premium? This partially depends on how frequently they get that SKU data. They disclose that the average active US customer transacts 5.4 times per year. In contrast, a typical debit-centric consumer generates ~400 transactions a year. I am not sure 5 observations allows too much pattern building, even with SKU.

On the positive side, 69% of those US purchases, (3.7 out of the 5.4) are in the “Apparel & Accessories” vertical, so less data goes a longer way – but is it enough for real predictive power? Certainly, Klarna has power users that transact a lot more, building deeper data sets, but, by definition they are a smaller subset of active users.

Finally, not all Klarna transactions get SKU.

• When Klarna is directly integrated into a web site, it does. This seems to be 91% of volume (page 120)

• When Klarna is not integrated it gets only the basic credit card data of amount and merchant. This happens in the following circumstances:

  • Physical Points of Sale with the Klarna Card or a digital wallet

  • Non-integrated eCommerce merchants (Klarna is at checkout but not in the purchase journey)

• SKU-capturing transactions are growing at 12% while non-SKU transactions are growing at 20%.

Deposit Strategy

“… we are well positioned to build a leading presence in strategic adjacencies such as retail banking services, given our trusted relationships with consumers, our experience and our existing banking services in select regions”

Klarna has a banking license in Europe that raises deposits already:

“..74% of which were fixed-term with an average duration of 280 days in 2024. As a result, we can adjust our lending policies quicker than our deposit base might change. We also have the ability to deliberately change the length and interest rate on the deposits that we offer to adjust this duration gap.”

If you have been following my posts, you know that I am not a big believer in US Neobanks, who bank the same population that Klarna lends to. Most of them lose money. Some of them have begun lending to their depositors, but the track record of entities starting with lending and moving to deposits is not encouraging.

So that one of my good friends doesn’t remind me of this yet again, Nubank is Brazil is a non-US example of doing this successfully. Point taken! But few US entities have managed the same trick. Both American Express and Discover have tried, but most of their deposits are still interest bearing at high rates. Synchrony and other monoline lenders mostly don’t even try for transaction accounts.

Capital One may be an exception, but they don’t publish enough data to know for sure. The other outlier is SoFi which bought a bank to get deposits, and that bank now has over $2B in DDA with $22B in interest bearing; but SoFi lends to a higher FICO population where DDA balances are also higher. Further in 2024, SoFi DDA deposits shrank by 11% while interest-bearing grew by 48%. As a rule, penetrating the US DDA market from a lend-centric customer base is not promising.

Yet, DDA deposits would be the way to lower cost-of-funds, not high-yield savings. High-yield savings are only a stable source of funding if you always pay near the top of the market. DDA deposits generally pay no interest and are more stable, but you need affluent consumers or small businesses to generate material balances. US Neobanks have neither. Klarna wouldn’t either.

Finally, all Klarna customers already have a DDA because they need to pay installments via debit cards. Unlike most Neobanks, which appeal to the unbanked or new-to-banking, Klarna has to get consumers to swap banks. That is very hard to do even for incumbents. The average DDA lasts for 15+ years.

Competition

Klarna was a pioneer of the Pay-in-4 model, and is the global leader, but they have competition:

• Major Pay-in-4 competitors:

  • Affirm has roughly the same US GMV, but most of its loans are longer term. Pay-in-4 is 15% of GMV but grew 23% in 2024

  • o AfterPay (Block) has global volume of $33B and grew 20%

• After checkout competitors:

  • PayPal reports $33B in 2024 global BNPL volume which grew 21% over 2023

  • Credit card-based solutions from Citi (Flex), Amex (Pay-it-Plan-it), and Chase (My-loan, My-line). These programs charge fees rather than interest but let the cardholder structure an individual card purchase as installments

  • Debit card-based solutions from FIS and others are just emerging

• Alternative borrowing options for credit-constrained consumers

  • Earned Wage Access (EWA) appeals to the same population for the same reasons and is a near substitute. It is generally free or low cost.

  • “Advances” from banks. These typically incur fees but repay automatically, timed to direct deposit inflows

Pay-in-4 is superior since it is in the commerce purchase path; the consumer sees the option while they are shopping rather than at checkout or afterwards. Pay-in-4 is also truly free to the borrower whereas others sometimes have fees. But this is a competitive space.

One major challenge is major retailers and platforms. Klarna recently displaced Affirm at Walmart. The problem with winning at Walmart is that the economics are usually not that attractive. Big retailers, including Shopify, Amazon & Apple Pay, often support two or more BNPL providers . Those merchants probably squeeze lender economics while relying on competition to give consumers a good deal.

This is no longer a white space arena. Not everyone can be underwritten at the low-end, and at the higher end, Credit Cards are simply a better mousetrap given rewards and grace periods. Pay-in-4 works for the credit-constrained – which is a significant TAM – but that TAM is fought over by a variety of providers.

Conclusion: Klarna doesn’t need the polemics

Klarna’s core business is sound. They get merchants to pay credit card-level MDRs so that credit-constrained consumers can level out their cash flow. Those consumers can then buy more – either more purchases or more expensive tickets. Merchants are willing to pay a premium to get that sales lift as there is a provable link between the lending and the lift.

Klarna has proven that its platform has marketing value – although it is unclear how much its data contributes to that. They are expanding into longer term, larger loans, albeit at credit card-like APRs.

Klarna’s prospects are bright, but the over-the-top, anti-incumbent rhetoric is mis-guided.

• From a merchant perspective they are not usually a low-cost acceptance option:

  • They don’t displace high-IC credit cards, but low-IC debit cards or low-IC credit cards

  • When customers pay with Klarna virtual cards, acceptance costs are just as expensive as high-IC credit cards but for lower spending customers

• For credit-constrained consumers, they are an unmitigated gift:

  • For the debit-centric, Pay-in-4 is usually the only borrowing option as they don’t have credit cards – and it is free

  • For those with low-line credit cards, they can borrow extra, for free and avoid adding to their card balance at high APRs

• For credit worthy consumers, Pay-in-four is largely irrelevant

  • Affluent transactors already get 30 days free float, which is more than Klarna offers. They also get rewards, which Klarna doesn’t offer

  • Prime borrowers get bigger lines for bigger purchases over longer terms. They can increasingly structure them as installments.

  • Klarna’s longer-term alternative for larger loans seems to be just as expensive as cards, albeit other terms are more forgiving

Key competitors are just as friendly to consumers, just as costly to merchants, and don’t feel the need to run down the incumbents. The polemics just aren’t necessary.