Framework for Building a Credit Startup

These days, credit product startups are experiencing a Golden Age. 

In recent years, hundreds (maybe thousands) of these startups have emerged, each pioneering innovative ways to provide affordable and accessible credit. There is now a seemingly endless stream of credit products, including personal loans, mortgages, point-of-sale loans, buy now pay later, auto loans, corporate credit cards, invoice factoring, and more. This growth is attributed to a wide range of factors, including a low-interest rate environment, venture capital, and the rapid expansion of mobile and internet penetration. 

Despite these startups’ offerings and the factors behind their growth, all of their products came from a single, framework:

1. Insight
2. Data
3. Risk
4. Economics
5. Scale

While the framework is simple, it is crucial that founders of credit product companies understand and execute this framework in order to build a successful company. 

Below, we’ll explore what this framework is so powerful and essential 

1. Insight: Finding Your “Key” Insight

The first step is to identify the “key” insight. 

Identifying and defining what is lacking or incorrect in the current credit industry is the most vital step when launching a credit startup. It serves as the foundation for every decision and action taken in the future. Founders must take note of this crucial step. 

Let’s examine some key insights that have aided founders in launching successful credit startups:

• Brex (Credit Cards for Startups) – Despite the fact that startups are not considered high-risk credit users because of their backing from venture capitalists, they have difficulty obtaining credit cards for their businesses without a personal guarantee. Brex identified this gap in the market and launched a credit product specifically designed for startups.

• Stilt (Loans for Immigrants) – The traditional financial system in the U.S. misprices or ignores immigrants and visa holders. This is mainly because these potential borrowers lack a U.S. credit history, even though they are financially responsible and deserve affordable, low-cost credit products. Stilt recognized this key insight, which led them to create loans specifically tailored to immigrants and visa holders.

2. Data: How Can You Validate Your Insight?

After identifying the key insight, the next step is to collect data. 

In the past, data collection had to be done manually, but now it is easier to source and compile larger amounts of useful data. This data will be used to develop a proposed solution to fix the problem identified by your insight. To explore this further, let’s look at the examples mentioned earlier.

• Brex – Brex found that using the cash available in a startup’s bank account as collateral was an effective way to demonstrate that startups are not high-risk credit users. By using this method to underwrite the line of credit, Brex was able to prove that startups with funding are not risky. This approach was particularly effective because startups often have a significant amount of VC cash on hand, which made proving their low-risk status relatively straightforward.

• Stilt – Stilt recognized that immigrants and visa holders represent lower-risk borrowers. However, traditional financial institutions relied on a borrower’s credit score to assess risk, which presented a challenge since most non-citizens lacked a credit history in the U.S. In response, Stilt relied on an alternative set of data that considered factors such as borrowers’ education, employment, and available international data sources to underwrite loans.

In both cases, these startups utilized unique data to address the issue of mispriced risk in underserved markets. 

It’s important to ensure that the data collection is scalable in the long run and remains stable and consistent at scale. It’s worth noting that the data collection methods of both Brex and Stilt can be replicated for any number of borrowers and can consistently prove their insights. 

While data collection may not always be the top priority, it should be considered while validating the initial hypothesis.

3. Risk: What Risk Model Can You Build from This Data?

After identifying your insight and collecting data to support it, the next step is to develop risk models and guidelines for your business. 

The first step in creating a lending risk model is to determine the purpose of the model. For instance, you may want to predict the likelihood of a borrower defaulting on their loan or identify the characteristics of a high-quality borrower.

Once you have determined the purpose of the model, you will need to return to your data. This involves selecting the variables that will be used to create the model. These variables may include borrower information, credit history, income, employment history, and other factors that may impact the borrower’s ability to repay the loan, including (and especially) the unique data you used to validate your key insight.

There are several modeling techniques that can be used to create a lending risk model, including logistic regression, decision trees, and neural networks. The choice of technique will depend on the purpose of the model and the complexity of the data set.

Once you have selected a modeling technique, you will need to train the model on the data set. This involves using a portion of the data set to build the model and then testing the model on a separate portion of the data set to evaluate its accuracy.

After training the model, you will need to evaluate its performance and refine it as necessary. This may involve adjusting the variables used in the model or changing the modeling technique.

Once you have developed a lending risk model that meets your needs, you can implement it into your lending operations. This may involve integrating the model into your loan application process or using it to assess the risk of existing loans in your portfolio.

It is important to monitor the performance of the model over time and update it as necessary to ensure that it remains accurate and effective. This may involve retraining the model with new data or adjusting the variables used in the model.

4. Performance: Can Unit Economics Work (in Theory)?

The model can only take you so far. To obtain real data, you must disburse actual loans. 

The number of loans and how quickly they are disbursed depends on the product and target market. To illustrate, payday loans have smaller amounts and shorter terms, while mortgages require larger sums and longer terms. 

Though the exact metrics to determine performance varies, early performance is crucial for generating loan performance. Unit economics is a method of analyzing a business model at a micro-level by evaluating the costs and revenues associated with a single unit of product or service. It helps a you to determine whether your product can generate profits on each unit sold and how much it would cost to acquire each customer.

By calculating the unit economics, you can evaluate whether its model is sustainable in the long term and profitable. It includes various metrics such as customer acquisition cost (CAC), lifetime value (LTV) of a customer, gross margin, churn rate, and others.

These metrics provide insight into how much your business is spending to acquire each customer, how much revenue a customer generates over time, and the overall profitability of the business. By understanding and optimizing unit economics, you can create a profitable and sustainable model.

If unit economics are poor, better pricing and underwriting models are necessary. Loan economics is crucial for credit startups as it is the only way to make money. Models improve with more data but the scope of growth is uncertain. A judgment call must be made based on the target market and potential spread.

5. Scale: Can You Grow Rapidly Without Material Changes?

To avoid data and underwriting issues as you grow, standardization is key. 

Once you have proven your business model at a small scale, it’s important to standardize your data collection and underwriting process. Standardization will ensure scalability or show you which parts of your model will break when scaling.  

This is also the right time to standardize other aspects of your company that we haven’t discussed yet, but will below.  

Investing in technology can automate processes and make the business more efficient – software, tools, and systems that can streamline operations and improve customer experience.

You will also need a strong team to scale. Hire talented and passionate employees who share the company’s vision and culture. The team should be able to manage growth, adapt to changes, and take on new challenges.

To scale, a startup needs to reach more customers. Focus on standardizing the marketing strategies that will attract a wider audience and improve customer retention.

Continue to focus on unit economics, too. It is essential to understand how much profit is made from each customer. Optimize pricing to improve margins and find ways to reduce the cost of acquiring and serving customers.

Managing cash flow is essential for a startup to survive and scale, too. Plan budgets and projections, and monitor finances closely to avoid running out of cash.

Scaling a startup requires long-term planning. Develop a roadmap for growth, set realistic goals, and measure progress. Stay agile and adaptable to pivot the business if necessary.

Finally, growth also means a larger focus on compliance because credit is a highly regulated industry. Credit products are highly scalable and predictable sources of revenue if data and underwriting models hold as you grow. While compliance is important, though, it’s key to focus on your insight, data, risk, performance, and scaling in the early stages. 

Final Thoughts

The framework for a successful credit start is simple but essential. If you want to succeed, burn into your memory:

• Insight
• Data
• Risk
• Performance
• Scaling

By identifying a key insight, validating it with data, creating a lending risk model, analyzing unit economics, and standardizing operations for scaling, credit startups can create a profitable and sustainable business model. 

With careful planning and execution, your credit product startups can continue to thrive and innovate in this Golden Age of credit.

Frank Gogol

A seasoned SEO expert, Frank has a long history of working with and for startups. Starting in mid-2018, Frank served as the SEO Strategist for Stilt, a fintech startup that provided fair loans for immigrants in the US and other underserved markets. While with the company, he scaled site traffic from zero to more than 1.5 million unique visits per month, driving the bulk of the company’s lead generation until it was acquired by J.G. Wentworth in December 2022. As employee #5 at Stilt, Frank was witness to, and part of, the successful building and sale of a fintech company, uniquely positioning him to create content for founders about all things startups.

See author's posts