Discover more from Rohit’s Newsletter
Lending business equation for your startup
If your startup expects to earn revenue from lending, here's how to think about your business model. I'll cover the most important revenue drivers, expenses, and their relationship as you scale.
All businesses can be distilled into an equation. Lending businesses are no different. In fact, lending is a deeply quantitative business.
Most lending products are commodities. Because of this, they are also super competitive. There is limited innovation (depending on how you think of the term), and all innovation is generally copied by the incumbents (i.e. banks) in due time.
Non-bank lending businesses are regulatorily and financially complex, have thin margins, are sensitive to macroeconomic changes, and require tight risk management.
The advantage is that both consumer and business lending are huge markets.
Just to put it in perspective, total non-housing consumer debt is $4.71 Trillion as of Q2 2023. Of which, $1.03 Trillion is credit card debt.
If we assume an 18% average interest rate, that’s $180 billion in annual revenue. Then there are fees, interchange, and other revenue sources. That’s a big market.
Lending P&L has a few important revenue and cost drivers. Let’s understand each of them, their importance, their relationship with other variables, and their importance at various growth stages of the business.
This is, generally, the biggest revenue driver. The interest is earned on the outstanding balance, not on the originated loans. Depending on your product, you may not earn interest revenue e.g. charge cards.
For charge cards, you don’t charge interest for 30 days and the balance is paid in full every month. So, for products like these, it’s important to figure out a business model with limited or no interest revenue.
Personal loans, auto loans, and mortgages start accruing interest on day 1 after origination.
For products like business loans, personal loans, auto loans, and mortgages - lenders earn a fee upfront for originating the loan (underwriting and funding the loan).
This fee is 5%-10% of the principal amount at the time of origination. It is lower for auto loans and mortgages. High quality borrowers are sensitive to this fee because it is paid by them and impacts the APR.
The fee is earned once at the time of origination and earned independently of repayments.
Origination fee is risk free revenue and the most important revenue component.
This revenue is for collecting the payments for the outstanding balances.
Depending on the type of product, the servicing fee is generally 0.25% to 1%.
Large balance loans like mortgages and auto loans have lower servicing fees. Higher quality credit with a high percentage of auto-pay setup will have lower servicing fees.
The fee doesn’t change as you scale and a smaller percentage of a lender’s revenue.
Referral or Merchant Fee:
These fees are earned if you refer the loan to others or help a merchant generate a sale.
Newer lending products like BNPL earn a fee from the merchant if the consumers complete the purchase using a loan. This has been true for auto loans for the longest time but now also applies to lower priced product purchases.
For some companies like Affirm, merchant fee is a material contributor to the overall revenue.
These fees include all other revenue from sources like late fees, insufficient funds fees, interchange fees (for credit cards), etc.
For some products, like payday loans, these fees add up and for most charge cards, interchange may be the only source of revenue.
Cost of Capital
This is the interest you pay to your debt investors. This is one of the biggest variable expenses for your company.
As you grow the business, the interest rate on your facilities will go down.
So you shouldn’t worry about the cost of capital in the earlier stages.
Cost of Acquisition
The cost per originated loan, credit card, etc. This cost is, mostly, the biggest expense for lending driven businesses.
When operating in a risk driven industry, managing CAC across channels is tricky. It increases exponentially and the quality of customers decreases rapidly. i.e. you’ll get a worse quality customer for the same CAC.
Cost for origination
This includes all the costs associated with converting a qualified applicant to a funded loan. Some people may consider data costs as a part of origination costs.
The human labor cost for origination is generally high in the early stages. Through iteration and learning, this comes down significantly with scale.
This is a constant cost driver for businesses like credit cards. As the business scales, the costs of rewards also increase at the same rate.
Rewards like points, cash back, etc are spent even if there is no revolving balance.
Credit card businesses (especially the ones that are paid in full every month) have negative unit economics because of rewards.
Some companies consider rewards cost as a part of their Cost of Acquisition.
This is the percentage of the outstanding balance that is not paid.
Defaults are a loss for the business, a big drain on unit economics, and may kill the business.
There are different types of fraud but I’m talking about fraud due to stolen information, synthetic identities, mass attacks, etc.
First party fraud i.e. consumer taking credit with the intention of not paying back is excluded.
This is the cost to underwrite one funded loan or credit card. It also includes cost of all the declines.
To underwrite a customer, you’ll need to pull data from multiple sources. As you grow, you’ll need to pull data from more vendors per customer. These costs add up. You pay for all applicants (including the ones you don’t conver). That’s why it’s important to maintain a healthy conversion rate.
Common sources include credit report, fraud vendors, credit scores, compliance check vendors, etc.
This is your cost to service the loans. These costs are mostly agents for customer service and processing payments.
Generally, it’s a small percentage (sub 1% annualized) of the outstanding portfolio balance. With automation and self serve options, it could be fairly small.
Thanks for reading Rohit’s Newsletter! Subscribe for free to receive new posts and support my work.
Here’s the equation you need to optimize:
+ Origination Fee
+ Servicing Fee
+ Referral/Merchant Fee
+ Other Fee
- Cost of Capital
- Cost of Acquisition
- Cost of Origination
- Rewards Cost
- Default Rate
- Fraud Cost
- Data Cost
- Servicing Cost
Only a few are important and only some of these variables are correlated with each other. As you scale the business, it’s important to get a couple of them really right.
Those variables are Interest, Cost of Acquisition, Default Rate, and Fraud Cost.
That’s pretty much it. The tighter your control on these variables, the stronger the business will be at scale.
At the early stage, the most important thing to focus on is your product’s value proposition. Until you have a good sense of whether your value prop is resonating, everything else won’t matter. After you validate the value proposition, generally focus on:
Cost of acquisition - It’s tough to acquire customers for fintech products. But if your fintech product is not spreading through word of mouth or because of your value prop, it would be a tough road ahead. CAC is generally low for fintech businesses in the early days. Then it rises rapidly.
Default rate - If you are lending and expecting to make money from lending, your default rate should be lower than expected in the early days. It doesn’t have to be zero but it should be lower than similar businesses. Default rates increase as you scale.
For early stage investors, at least these 2 metrics will be important. Fintech products are generally commodity products, so a low CAC and low default rate are important.
When scaling a lending business, it’s important to understand the inherent relationship between these variables. In some cases, these relationships are not linear and the risk increases exponentially. Mid-stage businesses are at the highest risk of things going wrong. When scaling a lending business (post Series B), keep a close eye on the following metrics:
Interest Rate - It turns out that if you lower your interest rates, you can get a ton of customers. And as soon as you increase them, customers will go away.
Cost of Acquisition - Scaling a lending fintech company is extremely competitive. As you focus on growth, make sure you are managing the risk. Growth will come either from optimizing and scaling existing channels or adding new ones. In lending, scaling existing channels changes the risk profile and every new acquisition channel brings its own set of problems.
Fraud Rate - The amount you lost upfront (at the time of acquisition) or post origination (due to intentional nonpayment) increases exponentially. High fraud rates can kill a fintech business. They are highly correlated with automation and channel risk.
Default Rate - When scaling acquisition, the risk profile of marginal applicants is different from your applicants at the early stages. The risk usually goes up, fraudsters start targeting your company, and default risk goes up exponentially. Keeping a close eye on upfront defaults and early signs of defaults for the new cohorts is extremely important.
Data Costs - The data required to limit default and fraud rates goes up (per unit) as you scale. You will need to integrate a lot of new APIs to filter fraud and your approval rates will go down (because of the increasing # of applications). With more API unit costs per application and more high risk applicants, data costs may spiral out of control. They can be controlled by managing your marketing channels.
If you keep a close eye on these metrics, you can scale profitably. Many companies start focusing on optimizing things like their cost of capital too soon. This is not advisable. It is a distraction and you may end up boxing yourselves in a credit box that doesn’t scale. If you have a capital markets team, they’ll continuously work on this anyway. If your business is working, the cost of capital will go down. Don’t stress about it. But if you get a great deal, go for it.
When you have built a big enough book of business, your main goal should be optimizing the right variables. At this stage, your cost of capital should be the lowest it can be. If you are still paying rates higher than competitors or your previous facilities, something’s not right.
Make sure that your mix of acquisition channels is delivering a stable CAC and stable quality of customers. Just follow a measured approach of growth in all the channels that work while testing new channels at the same time. The proven channels are your key to delivering sustainable risk adjusted returns.
Continue to drive down fraud costs through backtesting new algorithms. Optimize between great customer experience and fraud.
At this stage, you are constantly optimizing an equation with 10-12 variables for maximum profit.
Understand the relationship
As you scale, learn how these metrics work with each other.
The cost of acquisition is inversely proportional to net risk i.e. it is directly proportional to interest rate and inversely proportional to default cost and fraud rates. You may get lower CAC but it may end up costing on the other side in terms of poor risk performance.
If you charge higher interest rates, lower risk people will opt out of taking credit from you, and higher risk population will result in higher defaults.
At scale, pay close attention to how moving one metric impacts other numbers. It is a business of tradeoffs.
Below are a few sample P&Ls of lending driven companies. As you read through these S-1s, you’ll start to see all these variables depending on the company’s business model.
Because Affirm is a BNPL lender, its merchant revenue is the largest revenue driver. They don’t charge any origination fee to the customer.
Oportun is a sub prime lender to the Hispanic population and they don’t charge origination fees. Interest income is their primary revenue driver. Their cost of capital is comparatively low which results in high net interest margins.
Upstart started as a direct lender and tried to become a marketplace where banks originated loans using Upstart’s risk model and paid them a fees. Upstart also earned origination fee.
Lending Club started as a marketplace where they did not hold almost any credit risk for the originated loans. Their main source of income was origination fee.
**Lending Club is now a bank, so their recent financials may look different.
This should give you a fair overview of how to think about a lending business.
Hope it was helpful and feel free to reach out if you have any questions.
Thanks for reading Rohit’s Newsletter! Subscribe for free to receive new posts and support my work.