Klarna IPO - how should it be valued compared to Affirm?

Klarna is the next big fintech IPO and recently filed its F-1. Comparing Klarna's numbers to Affirm to determine a potential valuation. Affirm is similar, not the same but still serves as a guide.

Company Overview

Klarna is a Swedish fintech company founded in 2005 that pioneered the Buy Now, Pay Later (BNPL) space in Europe before expanding globally. The company's mission is to make shopping "smoooth" by offering flexible payment solutions that benefit both consumers and merchants.

Klarna was founded by three Stockholm School of Economics students: Sebastian Siemiatkowski (CEO), Niklas Adalberth, and Victor Jacobsson. Interestingly, when they pitched their BNPL idea at a startup competition in Stockholm, the judges rejected it. However, they managed to secure initial investment from Jane Walerud, who not only invested but also connected them with a team of programmers to build the initial product.

Product

Klarna's main products focus on flexible payment options for e-commerce:

1. Pay in 30 days: Customers can try before they buy and only pay for what they keep.

2. Pay in 3 or 4: Split payments into 3 or 4 interest-free installments.

3. Financing: Longer-term financing options for larger purchases.

4. One-click payments: Simplified checkout experience.

5. Shopping app: A comprehensive shopping experience with features like price drop notifications, collections, and delivery tracking.

Klarna integrates with merchants through APIs and provides a seamless checkout experience both online and in-store. Unlike traditional credit products, many of Klarna's payment options are interest-free for consumers, with the company earning revenue primarily from merchants.

Business Model

Klarna has several revenue streams:

1. Merchant fees: Similar to Affirm, Klarna charges merchants a percentage of each transaction (estimated 3-5% on average, higher than standard credit card processing fees). This is the primary revenue source.

2. Consumer interest and fees: For longer financing options, Klarna charges interest. In markets where permitted, they also charge late fees (though they've reduced these in recent years).

3. Banking services: In some markets, Klarna operates as a licensed bank and offers savings accounts and other financial services.

In 2024, Klarna reported revenue of ~$2.8 billion USD, representing a 21% increase from the previous year. The company's GMV (Gross Merchandise Volume) reached $92 billion in 2023, a 17% year-over-year growth.

Funding History and Valuation Trajectory

Klarna's valuation history has been volatile (given it's a 20-year-old company, and fintech has been volatile). COVID was particularly volatile for fintech companies that prioritized growth over profitability during the cheap-money era:

• 2007: Received first major investment from Investment AB Öresund, valuation around $60 million

• 2011: Sequoia Capital leads a funding round, valuation approximately $100 million

• 2015: Valuation reaches $2.25 billion following investment from Visa and others

• 2019: Valuation grows to $5.5 billion with backing from Dragoneer, Commonwealth Bank of Australia

• 2020: Valuation jumps to $10.65 billion (Silver Lake led round)

• 2021 (June): Peak valuation of $45.6 billion after SoftBank-led funding round of $639 million

• 2022 (July): Dramatic 85% devaluation to $6.7 billion in a down round ($800M raised)

• 2023: Valuation rebounds slightly to around $7.8 billion

• 2024: Current implied valuation is approximately $14.6 billion based on Chrysalis Investments' stake value

This valuation history mirrors the broader fintech funding environment: explosive growth during the 2020-2021 cheap money era, followed by a dramatic correction as interest rates rose and investors demanded profitability.

Major Investors

Klarna has attracted an impressive roster of investors throughout its journey:

• Sequoia Capital: Early investor that has participated in multiple rounds

• Visa: Strategic investment showing payment network interest in BNPL

• Atomico: Niklas Zennström's VC firm has been a longtime backer

• Silver Lake: Led 2020 funding round

• SoftBank Vision Fund: Led the 2021 round that valued Klarna at $45.6 billion

• Commonwealth Bank of Australia: Strategic investor

• Chrysalis Investments: Notable investor whose recent valuation mark-up has signaled confidence

• Mubadala Investment Company: Abu Dhabi sovereign wealth fund

• Dragoneer Investment Group: Participated in multiple rounds

Financials

Revenue

Klarna has demonstrated consistent revenue growth over recent quarters, with particularly strong performance in early 2024:

The company has maintained 20%+ year-over-year growth throughout 2023 and 2024, with Q1 2024 showing the strongest performance at 29% YoY growth. This consistent growth trajectory is impressive given the challenging macroeconomic environment and increased competition in the BNPL space.

Klarna has pursued growth across a few areas:

1. European expansion: Starting in Sweden, Klarna expanded across Europe, establishing itself as the dominant BNPL provider in the region.

2. US market entry: In 2015, Klarna entered the US market, facing fierce competition from Affirm, Afterpay (later acquired by Block), and eventually mainstream financial players like PayPal and Apple.

3. Merchant acquisition: Klarna has aggressively pursued integration with major retailers. They now work with over 675,000 merchants globally, including giants like H&M, IKEA, Samsung, and Nike.

4. App-centric strategy: In recent years, Klarna has shifted toward becoming a consumer shopping destination, not just a payment method at checkout. Their app has evolved into a comprehensive shopping platform.

The strategy has paid off: Klarna now processes 2.5 million transactions daily (based on the 2023 report) and has 93 million active consumers worldwide. Their customer retention metrics are impressive, with early cohort data showing purchase frequency increasing from 10x in year 2 to 20x by year 7.

Noteworthy - repeat usage

Klarna is telling a story of a growing number of users and higher usage amongst customers.

According to the SEC filing, Klarna reports the following about their user transaction frequency:

1. Year 1 (First Year) Usage: "On average, our 2019, 2020 and 2021 consumer cohorts made three transactions during their first year on our network"

2. Year 2 Usage: While the document doesn't explicitly state the exact number for year 2, it mentions that users reach "at least ten transactions by year three," suggesting a progression from 3 transactions in year 1 to approximately 6-7 transactions in year 2 (based on the growth trajectory to reach 10+ by year 3)

3. Year 3 Usage: "at least ten transactions by year three"

The frequency of usage is significantly higher in mature markets:

• Mature Markets: "In Sweden, for example, our average purchase frequency has reached 32 times per year in 2024"

• US Market: "In the United States, purchase frequency in 2024 was more than five times per year and grew 17% year over year"

Klarna is also accelerating Average Revenue Per User and Purchase Frequency in newer cohorts.

Klarna grew active users from 79M in 2022 to 93M in 2024 (17.7% growth).

Revenue Mix

Klarna divides its revenue between Transaction and Services Revenue and Interest Revenue.

Transaction Revenue = Merchant + Advertising + Customer Service

Interest Revenue = Interest Earned

Transaction revenue has marginally decreased over the last 3 years from 78% in 2022 to 75% in 2024.

This could be primarily due to US expansion and consumers choosing a longer-time financing product, which leads to higher interest income.

The same is true for Consumer Service Revenue, which has increased from 12% to 16% in the last three years. This implies that Klarna is charging more fees to consumers. The biggest fee is “Reminder Fees,” which represented 94% in 2022, 74% in 2023, and 83% in 2024.

US Market Expansion

One of the most significant stories in Klarna's recent performance is its accelerating growth in the US market:

The growth metrics are striking:

• US Revenue: 38% YoY growth in H1 2024

• US Gross Profit: 93% YoY growth in H1 2024

• US GMV: 50% YoY growth in Q3 2023

The substantially higher gross profit growth compared to revenue growth suggests that Klarna is not only expanding in the US but doing so with improving economics. This is crucial for the company's long-term position as the US represents both the largest potential market and the most competitive landscape for BNPL services.

Merchant Growth

Klarna shares the same story about merchant growth. Merchants have grown substantially over the last five years, and revenue per merchant is also growing.

The average revenue per merchant still grew by 6% for the ones onboarded in 2018. Recent cohorts show higher growth.

It’s not clear if Klarna is considering all onboarded merchants or only the active ones. I’m assuming that it’s all merchants in that cohort.

Portfolio Performance

Klarna has rapidly reduced its overall credit losses (as a percentage of GMV) over the past six years.

Klarna attributes this to a combination of developments in their underwriting models, product mix, and diversification.

• Real-Time, High-Frequency Underwriting: Each transaction receives a new underwriting decision in seconds, leveraging ML-enabled models that analyze proprietary data from ~2.9 million daily transactions.

• Small, Short-Term Exposures: Average balance per consumer of just $87 with 40-day average duration (versus $6,730 average US credit card balance and 2.9 years for typical US personal loans).

• Improving Model Accuracy: Klarna's Gini score (measure of predictive power) in the US improved from 0.36 in 2019 to 0.72 in H1 2024, significantly outperforming credit bureau models like VantageScore 4.0 (0.35).

• Broad Customer Base: Products serve consumers across the credit spectrum while maintaining portfolio quality through dynamic spending limits based on repayment history.

• Strategic Underwriting Adjustments: In H2 2022, Klarna implemented stricter underwriting standards including new risk models, increased down payments, and risk-based pricing to improve portfolio quality.

These efforts resulted in credit losses dropping from 9.6% to 1.2% in the US market and 0.47% of GMV in 2024 (overall), even as the company maintained growth.

Cost of Capital

Klarna is also actively changing its funding mix. It is moving towards a bank model and establishing a deposit-based funding approach. That provides a significant low-cost and long-term capital base, giving it an advantage over competitors.

Below are the key highlights from the SEC filing around the changing funding mix:

• Centralized Funding Model: Klarna Bank raises deposits and wholesale funding centrally, then distributes capital throughout the group, optimizing capital efficiency.

• Banking License Advantage: 94% of Klarna's lending in 2024 was funded through customer deposits, enabled by their banking license - a key competitive advantage.

• Deposit Control: 74% of deposits were fixed-term as of December 2024, allowing Klarna to actively manage the duration gap based on interest rate expectations.

Klarna is pushing for more funding through deposits but still maintains flexibility with other sources. Because Klarna’s average term is just 40 days, they can turn capital way faster than companies like Affirm (the avg loan term is 13 months).

Advantages of Klarna’s deposit base + funding mix:

• Duration Gap Creates Stability: The average term of deposits (280 days) versus the average loan term (40 days) creates a structural hedge against interest rate fluctuations.

• Flexible Funding Mix: While primarily deposit-based, Klarna maintains the ability to tap other funding sources including debt securities (investment grade BBB-/A-3 rating) and synthetic securitizations.

Path to Profitability

Klarna has faced pressure to demonstrate a sustainable business model, which they have responded to:

• 2019-2022: Significant operating losses

• 2023: Reduced annual losses by approx. $750 million, approaching breakeven

• 2024: Continued profitability in early quarters, strengthening its position for a potential IPO

Klarna's profitability comes from several measures:

1. Improved underwriting: Reducing credit losses through better risk models

2. Operational efficiency: Significant cost-cutting, including layoffs (about 10% of staff in 2022)

3. Higher margin products: Shifting focus to higher-margin services

4. US market growth: Scaling in the high-margin American market (US GMV up nearly 50% YoY in Q3 2023)

Total Operating Expenses as a % of revenue:

• 2022: 151%

• 2023: 114%

• 2024: 104%

Klarna has systematically reduced expenses even as it scaled active customers and merchants.

IPO and Current Valuation

Finally, Klarna filed for an IPO in the US market, with plans to raise approximately $1 billion.

Reports suggest the company is targeting a valuation of $15-20 billion, which would represent:

1. A significant recovery from its 2022 low of $6.7 billion

2. Still far below its peak 2021 valuation of $45.6 billion

3. A forward revenue multiple of roughly 5.35-7.4x based on 2024 revenue

Comparison with Affirm

Affirm is the closest comparison for Klarna, even though it has a slightly different business model.

Affirm’s primary revenue sources are merchant fees and interest, while Klarna’s 75% of revenue is just Merchant Fees (which they consider as Transaction Revenue).

Below is a comparison across multiple metrics:

• Affirm is smaller and growing faster, while Klarna is now profitable.

• Klarna is way bigger than Affirm with 90M+ active users while Affirm has 18M+ active users (21M per Feb 2025 news).

• Affirm’s revenue is ~8.7% of GMV, while Klarna’s revenue is ~2.4% of GMV

• Klarna primarily grew internationally before coming to the US, while Affirm is way bigger in the US and expanding internationally

Recently, Klarna has landed a few big customers like DoorDash and One App (Walmart).

Affirm’s market cap is ~$15 billion (as of March 21, 2025), and Klarna is seeking the same valuation.

It’ll be interesting to see how public investors value the company when it goes public in the next few weeks.

Finding regulatory orifices for fintech products

Building a fintech product is a delicate balance of - illegal and undifferentiated. It's easy to know if your product is differentiated, but how do you know if it's not illegal?

The biggest sin of a fintech company is an undifferentiated product.

The high cost of starting and running a fintech company, regulatory barriers, legal and compliance expenses, and other factors make it difficult for a fintech company to survive without differentiation.

Few fundamentally new products are launched in fintech, and even fewer become large companies.

I’ve realized that to build a large fintech company with a differentiated new product, you need to find a “regulatory orifice.”

What’s a regulatory orifice?

Regulatory orifices are tight gaps in the regulations that make it possible for your product to exist but have not yet been legally tested for your product. These regulatory gaps allow you to offer a new value proposition without violating the law.

There is no consensus on the legality of the product. At least 50% of the people you talk to will think it’s illegal. The remaining won’t be sure. Most “fintech pundits” despise these products and pronounce them to be non-compliant.

That’s where the real opportunity lies in fintech.

Why does it matter in fintech?

Very few large opportunities lie around in a highly regulated and contentious sector like fintech. This is especially true for banking and credit.

Most fintech products are a “commodity.” More and more fintech companies look alike.

As we have all seen:

• the next lender

• the next credit card

• the next checking account

• the next robo advisor

They all look the same.

Money is the definition of commodity, and most fintech products offer money in some form.

Without differentiation, it becomes a battle of CAC (which is, in any case, high for fintech).

Why do fintech products look the same?

Financial services are among the most regulated industries, and the U.S. is the most litigious country in the world.

External forces push fintech companies toward proven, legally-safe product designs. Imagine all the banks—when have you ever experienced a differentiated bank account?

External experts push for a safe product because fintech is probably the sector with the highest number of lawsuits.

Even positive innovation is met with strong pushback and extreme competition.

In other words, there are no positive points for helping someone; only negative points for not following the regulations.

Below are a few examples:

• District of Columbia AG filed a lawsuit against Earnin for deceptive marketing techniques

• In May 2024, the CFPB sued SoLo funds for falsely advertising no-interest loans. The attorneys general of Connecticut, DC, and Pennsylvania also sued for multiple reasons.

• FTC sued Dave (a banking app) in Nov 2024 for misrepresenting tips and false advertising.

• In Aug 2022, CFPB sued Hello Digit for lying to consumers about its automated savings algorithm

Multiple agencies have sued most household names, including Chime, Robinhood, SoFi, LendingClub, GreenSky, Goldman Sachs, and PayPal.

In fintech, getting sued is not a matter of if, but when.

You can (mostly) recover from lawsuits for false advertising, exaggerated claims, and misrepresentations, but one that makes the product illegal or forces you to change your product’s value proposition can.

That’s why it’s important not to be sued for your “product construct.”

That’s why lawyers and consultants are the most conservative on product definitions.

Following conventional legal guidance leads to building compliant but undifferentiated products that mirror existing solutions.

Everyone fears getting sued on the core product construct and having to shut it down completely.

A simple test

I have a simple test to determine if you are on to something:

If you find yourself pushing back and asking a lot of “why” questions, and all you hear is, “It’s how it is,” you know you are on the right path. You should feel confident about the product if they can’t point you to specific regulations or legal statutes that make it illegal.

With this test, you can also filter for truly creative and forward-looking experts. The good ones will lean in; the bad ones will just repeat what they have heard.

All regulatory orifices share these things:

• Unexploited gaps in the regulations to offer a unique value prop

• Misunderstood concepts broadly accepted as “best practices”

• Opened up because of new technology or change in consumer behavior

• Without a clear answer by industry practitioners

• Can only be figured out after building the product (but it gets there)

• Legally expensive to figure out (generally in a regulatory gray area)

What regulatory orifices are not:

• Going against established and clear violations of laws

• Previously tested and fined by regulators

• Gets your partners in trouble (either because they can’t support it or makes it illegal for them)

• Confidently considered illegal by truth-seekers in the industry

There’s no right answer, but I have found the above heuristic to be a good indicator.

A few examples:

• charging $10/month for a $250 line of credit (Brigit)

• no interest with an optional tip (Earnin)

• a debit card that helps you build credit (Extra app)

• offering accounts in the US without a physical US presence (Mercury, Stripe)

• lending based on school and GPA (Upstart, SoFi)

These value props sound like the norm today, and no one bats an eye. But when they launched, they all went through a trial by fire.

On the surface, these value propositions look like they partly violate a regulation.

Brigit* (and similar companies) are good examples of finding a regulatory orifice. These apps target low-income, low-FICO customers who are more likely to rely on payday loans.

Brigit charges a subscription fee to give consumers a cash advance of up to $250. The advance must be paid in full within one payment cycle (or the next payday). In theory, it’s like a payday loan, and accessing the loan costs $9.99 (subscription fee).

Regulators are cautious about products that put the most desperate consumers into debt. Many states including New York, New Jersey, Connecticut, and Massachusetts don’t allow payday loans.

Here’s guidance from CFPB on the definition of payday loans. According to the guidance, payday loans:

• Usually small dollar amounts ($100-$500).

• Relatively short term (2-4 weeks), usually on the payday

• Require repayment authorization

• Don’t need a credit check

Per this definition, Brigit’s original product would be considered a payday loan. Customers were charged $9.99 to access a $250 cash advance, which must be paid back at the next payment.

After extensive, expensive legal exploration, companies learned that a product that provides “valuable additional services” for the subscription fee is not a payday loan.

Because the subscription fee is not for the cash advance, it’s for the package of other services.

That’s why you’ll see all subscription plans have a ton of additional features/services like credit building, identity protection, rewards, financial planning, etc.

Consumers use these products primarily for cash advances (payday loans); however, because of the additional value-added services, they are not considered payday loans.

This is a “regulatory orifice” that consumer financial apps exploit, and it’s worth billions of dollars.

Another popular example is charging “tips” instead of interest or fees, which Earnin popularized.

Generally, lenders are required to show a Truth in Lending Disclosure (per Reg Z) summarizing the financial terms of the loan. For small-dollar loans, this APR could be in the hundreds of percent, making them unviable for most consumers.

But tips are voluntary, so they can’t be considered interest.

This interpretation allows the product to be widely available and not scrutinized.

However, according to multiple lawsuits, Earnin defaults on a minimum tip, which makes it more difficult to cash out again if borrowers decline to give tips.

Finding this regulatory orifice has generated billions of dollars of combined market cap for these companies (aside from the lawsuits).

How do you discover a regulatory orifice for your product?

The important thing about regulatory orifices is that you can’t start by figuring out an orifice and then build a product.

You always build the product first (with enough confidence) and then determine for certain whether it is a regulatory orifice.

It is always ex-post, rarely ex-ante, and it’s always to build a different product that customers value.

Here’s how it goes:

Founders start a company because they want to offer customers a unique product.

They start by building the product and validating their hypothesis.

Based on this validation, they raise their first round of funding.

Then, to fully build the product and offer it to a broader customer base, they talk to potential partners.

These partners could be bank sponsors, networks, processors, payment gateways, etc.

The partners, who have never seen this product, don’t quite understand whether it’s legal.

If the product has been seen before and is considered risky or illegal, it is immediately shut down. If it is not, this starts a discussion with the internal legal/compliance teams, external counsel, and consultants.

Sidenote: This discovery process costs a lot of money. That’s why many partners only engage with startups that have raised substantial funding.

Over time, it was discovered that the product could be legal because it is in a regulatory gray area.

The product is somewhat blessed after expensive conversations with all of the above.

The product inevitably changes to reduce regulatory risk but manages to keep its original shape (unique value proposition).

As the product grows and gains attention, it undergoes multiple rounds of regulatory review. If it survives, it becomes an industry standard.

Exploiting a "regulatory orifice" is the only way to build a differentiated product without being sued out of existence.

Important:

When the product value proposition is tested with external parties like lawyers or consultants, it is important to push to preserve as much of the original vision as possible.

Their incentive is to push the product to the standard norm because they are all risk averse and generally don’t have an incentive to offer advice that risky. They don’t want to be associated with a product that gets caught up in legal trouble.

Tips for navigating the path to a differentiated product

Founders must be keenly aware of the product they want to build and keep the consumer experience and value proposition sacrosanct.

You should be willing to go deeper than the presumed experts.

It’s easy to give in to legal and compliance pushback and slowly modify the product to hedge your risks. Even though that advice is generally correct, you have to ask a lot of “why” when building a new product that doesn’t exist.

If you stumble upon a value proposition loved by consumers and start working with external counsel, pick the right law firms, consultants, and experts to help you find long-term opportunities.

Incline towards external counsel who helps you find answers, not the ones who just cite “industry best practices.”

Finding believers and explorers (who are experts) is the best way to discover a “long-term regulatory orifice.”

Hope this post was helpful.

*Upbound recently acquired Brigit for up to $460M in cash and stock.

Using AI in Underwriting: Borrower Income Calculations

Income calculations are complicated when error tolerance is low. Increasingly, borrowers have multiple sources of non-standard income which further complicates. How can AI help fix this?

For lenders, calculating income is one of the most important but overlooked aspects of underwriting. Calculating income can take hours depending on the company and the type of customer. Some companies build up to 20 different income calculators. They use these different calculators based on who is applying for the loan. And they maintain updated versions of each calculator whenever they update their underwriting procedures.

It always sounds simple when a lending company states ratios like “debt to income” or sets income cut-offs in credit policy. But it’s complicated behind the scenes. That’s why, sometimes, portfolio performance varies significantly even with similar lending policies.

Credit risk and operations teams spend a lot of time getting the calculations right. When operations teams ask for documents in a piecemeal way, it also leads to a poor user experience.

The way an origination process works is:

• submit a loan application

• accept loan offer

• submit documents for verification

• loan officer reviews the documents, triangulates and calculates income, figures out they need more documents, and asks for more documents to get accurate income

• submit additional documents

• the document submission process repeats

This should be, ideally, a single-step document submission process. The loan officer should be able to calculate income and list all the missing documents the first time.

AI can help calculate accurate income in the first pass and reduce this back and forth.

This is only applicable if calculating accurate income is important for your credit underwriting processes. You can automate income calculations if you have a high tolerance for error or if you are conservative in your loan offers.

Real lenders focus on accurate income calculations. e.g. mortgages.

During mortgage applications, underwriters collect an insane amount of documentation to calculate your income and free cash flow accurately.

You must comply with federal income calculation guidance in loan products like mortgages because they only buy eligible loans.

You may not appreciate the nuances unless you spend years in the depths of income calculators. Building and maintaining standardized, unbiased, accurate, and fast income calculators takes significant effort.

Similarly, underwriting workflow is one of the biggest areas of innovation in SMB lending. Startups are helping reduce errors and increase the speed of underwriting. Automating this is even more critical with embedded lending.

Overall, this is a huge pain, and I think using AI could be a potential solution to this problem.

Unpacking Income Calculations

Let’s review why income calculations are time consuming and tricky to get right. We’ll use mortgage income calculation as an example for this exercise.

Almost all mortgage underwriting is still completed manually. Because of the detailed and complex nature of underwriting, applicants have to submit tons of documentation for income calculation (and verification).

Selling mortgages to Fannie has strict income calculation requirements. Based on Fannie Mae’s website, for originating compliant mortgages, lenders have to complete the following:

• Income Assessment

• Asset Assessment

• Credit Assessment

• Liability Assessment

In this post, we’ll focus on only income assessment.

Assessing income for mortgage origination has a few components:

• Income sources

• Document verification

• Income calculation

• Income stability

In certain cases, lenders are required to consider income sources from many sources:

• Base income (W2)

• Bonuses (W2)

• Hourly (1099)

• Commissions (1099)

• Overtime (1099 or W2)

• Gig economy (1099)

• Military income

• Tips (1099)

• Government assistance

• Alimony or Child Support

• Cash deposits

• Venmo, Zelle, etc

• Rental income

• Dividend income

• Interest income

• Capital gains

• Foreign income (sources that are outside the US)

• Social Security

• VA benefits

• Retirement income

• Pension benefits

• Seasonal income

These are the majority of the types of income that could be submitted by the applicant.

Given applicants are increasingly working multiple jobs, there’s a lot more to parse through in their loan applications. With household income, income sources are exponentially more complex to calculate.

There is also variability in the frequency and consistency of these income sources. Below are a few examples:

• Bonuses could be paid every month, once a quarter, or annually

• Overtime could be inconsistent every month

• Gig economy workers could be paid daily, weekly, bi-weekly, bi-monthly, or monthly

• Hourly workers could receive bonuses based on hitting certain milestones

• Some positions may offer sales commissions which could be paid infrequently, like monthly or quarterly

• Rental income (or from sites like Airbnb) could be seasonal

Each income source can be proved with multiple types of documents. Based on your credit policy and underwriting procedures, you have to collect documents for all these income sources. It could quickly become complex once you consider all the sources of income and the ways they could be paid.

The applicant may switch jobs or work with multiple companies, so the applicant needs to submit this information from all those sources.

Depending on the income type, lenders may need to collect documents going as far back as 24 months. One can easily imagine the work needed to collect and verify these income documents. The amount of documentation needed to verify income sources across multiple bank accounts could be a nightmare for even the most experienced lenders.

Operational procedures could be complicated because of the significant variability in these documents. New hires need to be trained in these complex verification procedures as your team grows. Balancing customer experience, processing timeline, accuracy, and fraud gets tricky.

Fannie Mae has given guidelines on documentation needed for each type of income. Below are a few examples:

1. for commission income, they suggest collecting the last 24 months, but 12 months could also suffice as long as there are positive factors that compensate for a shorter history

2. for secondary employment income, different employers are acceptable as long as income was consistently received without more than 1 month gap across jobs in the last 12 months

3. for self employed income (where the applicant owns more than 25% of the business), they need to submit more documentation per the self employed documentation requirements

4. rental income is acceptable if it is established that it’ll continue, if yes, then form 1007 or 1025 (as applicable), and either most recent federal tax returns with Schedule 1 and Schedule E or copies of the current lease agreements

After you determine all the income sources, collect documents, and verify them, you have to clean the data and calculate income. Each type of income needs to be annualized and then evaluated for stability, consistency, and predictability.

Lenders are asking a few underlying questions:

• What's the pre-tax or post-tax monthly income?

• What's the free cash flow?

• What's the stability of income?

• What is the predictability of income?

Extracting and cleaning the data has its own set of challenges. OCR tools help make things easier and faster, but it’s still time consuming to check completion, validate documents, and verify outputs.

With this context, I will discuss how AI can help improve income calculations.

Automating income calculations for consumers

In consumer lending, accurate income calculations can mean the difference between a successful loan repayment and a default. Collectively, it can mean millions of dollars in losses. Being too restrictive in tolerance and calculations can also cost hundreds of millions in loan originations or adverse selection. The right balance between accuracy of income calculations and customer experience is core to building a successful loan portfolio.

Even with all the advanced tools around income calculation, a large percentage of income calculations are done manually. There are 4 steps to calculating income correctly.

1. Collecting documents

2. Verifying data

3. Cleaning inputs

4. Calculating income

AI can help with each of these steps. It can help reduce friction in collecting information from applications, increase accuracy and speed of document verification, help find areas of missing information, clean inputs, and finally, automate income calculations.

Let’s see how AI can help with each of these steps.

Collecing documents

AI can help collect complete and accurate documents in real time. Fine tuned models can process uploaded documents and provide feedback based on missing information.

AI can:

• generate automated emails requesting missing documents.

• highlight incomplete information or mismatches in uploaded documents

• detect duplicate document uploads

• check whether uploaded documents meet specific criteria, such as being within a recent date range or containing required signatures, and request re-uploads if necessary

These are some of the benefits of improving document collection from applicants.

Verifying data

Given we expect growth in fake documents, credit policies with become tighter. AI can help us verify information faster and reduce fraud. AI tools can automate most of the matching work while underwriters review the final output and sign off on the output. In verifying data, AI can:

• assess documents for authenticity by identifying indicators of manipulation, such as unusual fonts, pixel inconsistencies, or incorrect metadata, further reducing the likelihood of fraudulent submissions

• give underwriters a clear, concise view of each applicant’s profile, highlighting any potential red flags or areas requiring further investigation

• apply policy cut offs and ask for more documents as needed e.g. if 4 most recent paystubs are needed and the applicant uploads only 3, AI can give feedback in real time

• cross-reference information across multiple documents (e.g., comparing tax returns with bank statements) to ensure consistency and flag any discrepancies for review.

• automatically classify documents by type (e.g., W-2, bank statement, tax return), making them easy to locate and reducing manual sorting

Real time verification of uploaded documents could be one of the biggest benefits of using AI. This goes beyond companies providing document OCRs.

Cleaning data

Uploaded documents are incredibly messy. Until you have seen the wide range of information uploaded by applicants when applying for loan, you don’t comprehend the benefits of automation. The clarity of your instructions doesn’t match the quality of uploaded information. Getting clean data to work with is one of the biggest time savers for underwriters. For this area, AI can:

• extract key pieces of information depending on the type of document; because AI can classify the document, it knows how to extract the right pieces of data from the document

• combine data from multiple documents to build a clear picture for the underwriters

• translate documents in the right language if the uploaded documents are in different languages

• convert different payment frequencies (e.g., biweekly or weekly) into a standardized monthly income, providing an accurate, comparable view of income across applicants.

• clean and standardize applicant details (e.g., formatting names, addresses, and dates consistently), ensuring data accuracy and reducing confusion due to typographical errors or varying formats

• automatically check for typical errors or inconsistencies (e.g., incorrect Social Security numbers, invalid dates, or mismatched names across documents) and flag these for correction, improving the data’s reliability

Accurately extracting the correct information from documents is a time consuming task. Even though there are tools available in the market, we still have a long way to go towards automating data cleaning for underwriters.

Calculating income

Finally, AI tools can help put all this together, and the output can be used to calculate incomes. It’s relatively straightforward.

Incomes from various sources are annualized and discounts are applied based on consistency, sustainability, and predictability.

e.g. rental incomes may be discounted by 50% if they are recent or have been fluctuating over the past 2 years

To annualize incomes, dates and income amounts should have be extracted correctly. After this, the system calculates the income as needed.

Then all the income sources are added to get the final number.

There are standard income calculators after you have extracted the information. Here’s the Fannie Mae income calculator.

By using AI in all of the above steps, I think lenders can reduce income calculation costs and time by 50% - 75% (if not more).

Hope this helps you get a better overview of how AI is transforming operations at fintech companies.

Unpacking Embedded Lending: Components, Risk, and Business Models

I break down different types of embedded lending business models with their pros and cons. Choosing the right type of model can make all the difference in your fundraise, valuation, and success.

Embedded lending has made a comeback and gained significant traction in the past few years, especially in consumer fintech.

Auto loans originated at dealerships have been one of the most successful embedded lending products in history. Consumers like convenience, dealerships sell more cars, and lenders get access to borrowers. It was a win-win-win.

During the internet boom, companies like BillMeLater (originally called I4 Commerce) launched to take this concept online and offered simple credit financing for online purchases. BillMeLater was eventually acquired by PayPal in 2008.

There was a lull for a few years when Affirm launched in 2012. Buy Now Pay Later (BNPL) made a comeback around this time.

Fast forward to today, a few top embedded consumer finance companies originate $150+ billion annually and have generated a market cap of $30+ billion.

SMB embedded lending has also gained traction recently. Companies like Lendio, Settle, Parafin, and many others have raised significant capital.

Even with this growth, it seems like embedded lending is not understood that well. It sounds like a monolith but there’s a lot more under the surface.

So I thought I’d unpack embedded lending platforms, their components, and business models for founders.

Categories of embedded lending businesses

All embedded lending companies can be categorized across 3 dimensions:

• Target Market - Consumer, SMB

• Business Model - BNPL, Unsecured Loans, Secured Loans, Credit Cards, Invoice Factoring, Merchant Cash Advance, Working Capital, etc.

• Vertical - General, Healthcare, Retail, Auto, E-commerce, Real estate, SaaS, Services, Wealth Managment, etc.

The below is just a small sample from hundreds of embedded lending companies.

• Consumer/BNPL/Ecommerce: Affirm, Klarna, Afterpay

• SMB/Working Capital/General: Parafin, Kanmon, OatFi

• SMB/Merchant Cash Advance/Ecommerce: Stripe Capital, Shopify Capital

• Consumer/Unsecured Loans/Healthcare: Sunbit, Cherry, PayZen

• SMB/Invoice Factoring/General: Pipe, Wisetack, Slope, Resolve, Mesa

• Consumer/Credit Card/General: Imprint, Synchrony, Marcus by Goldman Sachs

We’ll look under the hood of a few embedded lending companies. We’ll learn about the components of a business, revenue drivers, pros, and cons. Being aware of different business models can help founders focus on the important problems applicable to their company, help them make the right decisions, and avoid going in the wrong direction.

After reading this, you’ll be able to answer questions like:

• How should I think about the right business model when I start?

• Do I need to raise debt capital for my embedded lending company?

• What are the pros and cons of various embedded finance business models?

Components of embedded lending businesses

Every lending company and by extension, embedded lending company has a few core components. We’ll quickly recap those below and then review how different companies make money.

If you are interested in learning about these in detail, feel free to read these posts:

• A brief guide to building your first lending product

• Building large embedded lending businesses

• Lending business equation for your company

Now, let’s list all the components of an embedded lending business. You can break down an embedded lending business into 5 components:

• Platform

  • The website that sells the product or provides a service.

• Originator

  • The embedded lending company.

• Bank

  • The bank sponsor.

• Debt Facility Provider

  • The debt facility provider.

• Loan Product

  • Installment loan, Pay in 4, Invoice Factoring, Line of Credit, etc.

All business models across all categories of embedded lending depend on the choices of these 5 components. Revenue, cost, team, partnerships, etc. are driven by your preference for each of these components.

Business models in embedded lending

I’ve previously written about building your lending business equation.

You solidify your revenue driver and business model based on your category and choice of embedded lending components.

You will see that banks like to earn interest while non-bank lenders prefer to earn fees. Generally, fee revenue is valued higher than interest revenue.

Some companies fight against fees for consumers and pass those costs to merchants, while others charge no fees but higher rates.

There are at least 4 different types of business models based on how you set it up.

• Consumer Embedded Lending

• Embedded Loan Marketplace

• SMB Embedded Lending

• Co-branded Products

There are others like Captive Embedded Lending but we’ll focus only on the 4 most common for this post.

Model 1: Consumer Embedded Lending

Consumer lenders acquire customers only by embedding their loan products in other websites. They have $0 CAC. These financial services companies generally originate a loan using a bank sponsor or lending licenses. They raise debt facilities (from banks or non-bank lenders) and perform all servicing/collection activities (sometimes they pass it on to third-party servicers).

Company: Affirm

• Platform

  • Peloton

• Originator

  • Affirm

• Bank

  • Celtic, Webbank, Cross River Bank

• Debt Facility Provider

  • Multiple large debt facility providers like Ares, Goldman Sachs, etc.

• Loan Product

  • Installment Loans

Peloton sells products (bikes) and wants to increase sales by offering credit to purchase their bikes. Affirm integrates with Peloton to allow customers to buy products using a BNPL product (close end unsecured personal loan).

Peloton launched with Affirm back in 2015. By 2020, 30% of Affirm’s originations were from Peloton.

Here’s how it works:

• Affirm integrates with Peloton to verify the delivery of the product, honor return policies, and process refunds.

• Affirm uses multiple bank sponsors like Celtic, Cross River, and Webbank for loan origination.

• This means that the credit policy is owned by the bank, but developed in partnership with Affirm. It’s Affirm’s credit policy approved by the bank.

• The originated loans are sold 3 days later back to Affirm. The originating banks may or may not participate in the originated loans. Generally, they keep 5% of the originated pool on their books.

• Affirm buys those loans using debt facilities from third party lenders.

• These are asset backed facilities and the lenders take most of the economics.

Affirm earns merchant discount rate and residual interest but also bears the risk of defaults.

In this model, Affirm is a lending company with $0 CAC. Once they are integrated with the store or service provider, acquiring each incremental customer costs $0.

Pros:

• Control over approval rates and credit policy.

• Can continue to scale with a few large debt investors.

• Can securitize and make good economics if loans perform.

Cons:

• Managing credit risk, cost of capital, servicing/collection teams, etc.

• Increasing interest rates could negatively impact unit economics.

• Manage bank sponsorships and slow iteration (everything has to be approved by the bank).

Model 2: Embedded Loan Marketplace

Loan marketplaces don’t take credit risk or build their credit policy. They don’t have bank sponsors, they don’t raise debt capital, or service the loans. They implement the lender’s credit policy in their system. Then they securely get the applicant’s information and either send that as a lead to multiple lenders (based on high level policy cut offs) or apply the credit policy to show pre-approved offers (similar to loan comparison websites).

Company: Lendio

• Platform

  • Lendio

• Originator

  • N/A

• Bank

  • 75+ lenders on the marketplace

• Debt Facility Provider

  • Banks or non-bank lenders

• Loan Product

  • Multiple products like Working Capital Loans, Term Loans, Credit Lines, ec.

Let’s take an example of a partnership between Staples and Lendio. Staples has tens of thousands of small businesses and they want to provide financial products to those SMBs. So, they partner with Lendio to offer these products.

• When a small business on Staples needs a loan, they can apply for a loan through Staples.

• The application flow and loan processing are from Lendio.

• As Lendio receives the application, they apply the credit policy of 75+ lenders on the other side and show offers to the small business.

• The business moves forward with one of the offers.

• The bank or non-bank lender whose offer was accepted processes the completed application.

• Lendio’s customer service helps the business throughout the process.

• Once the loan is finalized, it is “originated” and “funded” by the lender (bank or non-bank).

• Lendio earns a fee and doesn’t take any credit risk.

• The lender also continues to service the loan.

Affiliate fee is the primary revenue source in a marketplace business model.

You may have to share this fee with the Platform (Staples in this case).

Your net fee should be high enough to pay for: setting up partnerships (both sides), credit policy management, compliance, and customer support.

Different lenders pay different fees based on the quality of applicants and lending products.

Pros:

• Originator “marketplace” doesn’t take any credit risk.

• Needs a small team - primarily for partnerships, compliance, and customer service.

• No post origination responsibilities like servicing, collections, etc.

Cons:

• You don’t control the credit policy. Lenders may frequently update credit policies and rates.

• You have to manage policy updates which impacts approval rates.

• Generally doesn’t work for high risk populations (unless there’s also a turndown program).

Model 3: SMB Embedded Lending

Similar to consumer, embedded SMB lenders provide financing to businesses through a partner. Embedded lenders develop a credit policy, use bank sponsors, raise debt facilities, and perform servicing/collections. SMB lenders generally offer multiple products and more leeway in pricing (due to limited SMB regulations). Based on pricing, they may earn an origination fee and net interest margin.

Company: Parafin

• Platform

  • Doordash

• Originator

  • Parafin

• Bank

  • N/A

• Debt Facility Provider

  • Multiple large debt facility providers like SVB, Jeffries, Trinity Capital, etc.

• Loan Product

  • Merchant Cash Advance, Working Capital Loans

Parafin launched with Doordash as its first partner on Feb 9, 2022. Doordash announced “Doordash Capital” as their first financial product powered by Parafin. It’s important that Doordash retains the branding because Parafin could be replaced in the future. The first product is “revenue based financing.”

• DoorDash has a ton of suppliers, restaurants, and other small businesses. Parafin integrates with DoorDash and offers loans to those SMBs.

• Parafin controls the credit policy, probably with DoorDash’s inputs.

• DoorDash doesn’t take any credit risk and likely controls the customer experience.

• To the SMB, it seems like they are taking a loan from DoorDash.

• Behind the scenes, Parafin provides “merchant cash advance” financing i.e. Parafin gets paid as a % of sales and uses the debt facilities to fund the loans. Parafin then services and collects the loan.

• Doordash may earn a commission based on the loans originated through its platform.

Pros:

• Control over credit policy to only originate loans that meet profitability standards.

• Potentially high margin given the effectively high interest rate of the product.

Cons:

• Have to balance credit risk, approval rates, and net interest margins. High cost of capital may impact unit economics negatively.

• Need to constantly raise capital for first loss positions and grow teams to service the loans.

Model 4: Co-branded Products

Credit cards are the most common co-branded fintech products. Both for consumers and businesses. In a co-branded relationship, the brand provides distribution and the embedded finance company handles everything else - credit risk, debt facilities, card products, fraud, servicing, collections, etc. Co-branded relationships could be extremely valuable both for the brand and the embedded player. American Express and Delta is one of the most successful co-branded relationships (all large airlines have similar relationships). Even Apple built a co-branded card with Goldman Sachs (which wasn’t as successful).

Company: Imprint

• Platform

  • Brooks Brothers

• Originator

  • Imprint

• Bank

  • First Electronic Bank

• Debt Facility Provider

  • Citibank, Atalaya, SVB, etc.

• Loan Product

  • Credit Card

Brooks Brothers launched its first co-branded credit card with Imprint in July 2024. This partnership reflects the growing interest in monetizing through embedded fintech products.

• Brooks Brothers is a 200+ year old popular high-end brand.

• They have millions of customers and they want to offer products to increase loyalty. They partnered with Imprint to launch a co-branded card.

• The card is operated and managed by Imprint *in partnership with Brooks Brothers*. Imprint manages the credit policy, pricing, approval rates, line management, servicing, collections, etc.

• Co-branded cards have to offer strong incentives for customers to use branded cards as their primary card over bank cards like Chase.

• That’s why co-branded cards have competitive rewards with special rewards for usage at the brand’s stores. Imprint has to closely manage profitability because of these incentives (in addition to overall credit risk).

Imprint’s revenue driver is the interchange and interest earned. It has to balance that with the commission paid to Brook’s Brothers, default rates, and the rest of the operations.

Pros:

• Strong loyalty because of brand affiliation

• High LTV

• Zero CAC

Cons:

• Managing credit lines, interest rates, rewards, cost of capital, and risk.

• Need bigger teams for customer support and servicing.

• Complex accounting and compliance management across multiple partners.

Examples of additional embedded lending models:

Stripe Capital, Shopify Capital, Square Capital, and Toast Capital are examples of business models like captive embedded lending. In this case, the platform is the originator and lender (with or without a bank sponsor). But lending is a different product than their core product. They use lending as a monetization product.

Car manufacturers drive sales with 0% APR financing offers for new car purchases.

New home developers offer interest rates and cash discounts when taking mortgages with their preferred lenders.

Hope this post helped you understand the details of embedded lending.

Where does value accrue in public fintech companies?

We think of fintech as a giant monolith but it has many sub-verticals like lending, payments, banking, insurance, etc. Let's have a look at the winning and losing sector(s) in the public market.

Recently, I was taking a deeper look at “what really drives value creation in fintech” for founders and investors.

My theory has been that public markets are tough on pure fintech companies.

Private markets valuations are still dissociated with how companies will be valued when/if they go public. Acquisitions by entrenched players is the most common exit path (also true for everything in tech). But I still wonder about the multiples of these acquisitions.

Public markets are my best window into potential multiples for these deals (particularly late stage).

So I decided to look into how various fintech verticals have performed in the public markets.

If you are looking to start a fintech company, it’s worth paying attention to the sectors that have produced big exits. Public market performance is a big factor for venture investors when evaluating potential investments. This will help you guide your expectations on multiples and potential exit options.

And I used F-Prime Fintech Index* for a list of public fintech companies.

According to the index, fitnech companies have outperformed major indexes and ETFs (EMCLOUD, NASDAQ, and S&P).

• Fintech Index: +694.9%

• EMCloud Index: +418.9%

• NASDAQ: +271.1%

• S&P 500: +172.9%

Fintech Index has 51 companies across 7 sectors (Insurance, Payments, Banking, Lending, B2B SaaS, Wealth/Asset Mgmt, PropTech).

Let’s look at the performance of each of these sectors:

Insurance: -24.9% (6 companies)

Payments: +1,017% (19 companies)

Banking: +63.5% (6 companies)

Lending: -52.4% (4 companies)

B2B SaaS: +320.8% (8 companies)

Wealth/Asset Mgmt: -0.7% (4 companies)

PropTech: -55.8% (4 companies)

Only 2 sectors have done better than the market. Really one sector.

And that’s Payments.

In this exploration, I am obviously not looking at the performance of venture investors and valuations of private companies. Many are doing well that are not yet public.

But public market performance drives how late stage private companies are valued.

In fintech, there’s a clear separation of public market performance by sector.

Payments has the most number of companies (19) and the only sector that continues to outperform.

Payments outperformance is outsized because some companies became fintech companies later.

e.g. Shopify. They didn’t do payments for a while and the investment wasn’t led by fintech partners.

MecardoLibre - started in 1999, is an e-commerce company, and recently considered a fintech.

These 2 companies are the main reason that Payments sector is up 1,000%+.

If you take these 2 companies out, the performance would be significantly lower. It’s totally fine to have a couple of winners in a portfolio but these 2 massive winners were not originally fintech companies.

Nubank has been a massive outcome for venture investors but hasn’t *yet* performed great in the public markets.

Having said that, the last few years have been volatile (particularly for banking and lending companies). Maybe they’ll bounce back in the next few years with rate cuts.

Worth mentioning that many private fintech companies have high valuations (who raised during the pandemic), but only a few of them have raised at reset valuations in 2023 and 2024.

Most companies have also been payments companies like Stripe (at $70B) and Ramp (at $7.65B). e.g. I can’t remember any banking company that raised a round recently. Aven is close with their raise at a $1B valuation.

The question I ask myself: where does the value accrue in fintech?

At least in the public markets, using fintech for monetization has worked better than being a “pure fintech” company.

Probably that’s why “enabling fintech monetization” has been on of the key investment areas for VCs. This takes the form of B2B SaaS.

Most VCs have pulled back from investing in consumer fintech (particularly banking and lending) and it makes sense.

The bar for consumer products is way higher than before.

Most of the current public B2C banking/lending companies that were started in 10+ years ago are thriving. Below are a few big companies and their start year:

• Affirm - 2012

• Nubank - 2013

• Chime - 2012

• Robinhood - 2013

• Coinbase - 2012

• Opendoor - 2014

• SoFi - 2011

• Upstart - 2012

• Monzo - 2015

• Revolut - 2014

Founders look at big consumer fintech companies and want to start another one. But they need to think deeply about their differentiation. You can’t expect to build a large company by being similar to companies that were started 10 years ago and grew with the market. You need to figure out your own trend in today’s world.

And when you launch a new product, you need to reach escape velocity quickly.

B2B SaaS is also an interesting story, most of these companies were started 15 years ago. It tells you the time it takes to build a business a B2B company in fintech. There are 8 companies in the index, they were all started at least 15 years ago. A few of them started 24 years ago.

• Envestnet - 1999

• Xero - 2006

• Q2 Holdings - 2004

• Blackline - 2001

• Bill.com - 2006

• Alkami Technology - 2009

• Intapp Inc. - 2000

• Expensify - 2008

No company is currently $10B+ market cap.

Private companies are certainly worth a lot more. But public market comps are not generous.

Private market investors have made great returns investing early in these companies.

When starting a fintech company, know that it’s a long haul and you have to build with better fundamentals than other industries. It’s not the easiest sector and public markets are tough.

*The index is focused on startups that went public. It excludes companies like Visa, Mastercard, etc. Also includes all the banks.