The lending opportunity in vertical software:
“Somebody loan me a dime,
I need to call my old time, used to be.
Somebody loan me a dime,
Mmm, I need to call my old time, used to be.
Oooo, little girl’s been gone so long,
You know it’s worryin me.
Hey, it’s worryin, worryin me.”
Vertical software has been a consistent theme in Europe the last 5+ years. I was fortunate enough to witness from the sidelines new companies such as Mews being built in my previous role at Columbia Lake. Mews started with core software for hotel operations and layered in payments thereafter in 2 waves; firstly for bookings and secondly moving into restaurant payments via its acquisition of Bizzon last year. This is a well-known playbook pioneered by early vertical software winners in the US such as Mindbody who also managed payments (alongside SaaS) for its customers such as gym and yoga studios. Payments significantly increases the total addressable market in vertical software, which can seem small at first for many of these companies. By the time Mindbody went public in 2014 it was generating $4.3 billion in payments volume and $80 million in revenue (at 60% blended gross margin). By 2018, this had grown to over $7 billion in payments volume and $200 million in revenue (of which 40% related to payments and 60% to software licenses). Payments was and is still today a key revenue opportunity for vertical software that scales nicely (and without massive incremental cost despite being lower margin). It creates an option on a whole new business line that simultaneously compliments the core product leading to improved retention over time (try leaving Mews or Mindbody if all of your daily customer payments are managed through their systems vs. just being a software customer; it’s harder).
That said I believe there are other options for vertical software companies to maintain growth and move into new business lines that expand their markets further over time. Perhaps the best option I have been thinking about more and more is lending. Business loan options in the US and Europe involve loan processes with local banks that can be lengthy and painful, with documents flying back and forth over email and banks taking time to understand the business before making a decision. Moreover local banks will often ask for a personal guarantee for loans that are not secured over collateral or for smaller customers (source); thus borrowers such as local gyms or yoga studios (in Mindbody’s case) without real assets would find it harder to access loans for refurbishment or capital expenditure improvements than other businesses with real collateral.
Having worked for a bank whose primary business was making loans in the past, there tends to be 3 things that you need to make lending work well, especially if lending is the core of what you are building (i.e. you are not becoming a bank/holding deposits etc):
Good credit underwriting (people teams and/or models) driven by high quality data
Unit of value - how does the size/term of loan vs. distribution model/cost/sales cycle time work
Good credit monitoring/data sources from borrowers
For the first requirement, it’s easier to underwrite loans based on historical data provided by customers. It’s much better to combine this with third-party data. Hypothetically, the crème de la crème of lending would be to have your own primary data that accurately depicts historical financial performance and credit-worthiness of the potential customer before they even apply for a loan.
For the second requirement, it’s very hard to scale a lending business if you have relationship managers involved in the sales process and your loans are small and/or are short duration (<1-2 years). Again, this assumes you are not cross-selling other services like banking/FX which can increase your revenue footprint per customer. Lending is a lower margin business in general and so finding a cost-effective distribution channel that can scale well is table stakes for success. Thus if you could just have a slim team working through approvals, portfolio analysis and customer service without having to sell (i.e. find loan customers) that would be crème de la crème.
For the third requirement, having good performance data once a loan has been funded is very important. If you can get this regularly from a customer that is good but if you could get this weekly and it’s already in your own system (i.e. you don’t need to wait for the customer to send it to you) then that would be the crème de la crème for portfolio monitoring. You would have a real-time view of borrower performance across the loan portfolio and less administrative cost/time in requesting reporting from borrowers.
Now, applying this lending framework to vertical software companies demonstrates that vertical software companies get a crème de la crème in every department. Using my previous Mindbody and Mews example, both companies:
Own the historical customer performance data, best viewed through the payments (i.e. revenue) they take on behalf of customers. Even if no payments were managed by the software, lending could still be feasibly launched if KPI data on borrower performance existed in one’s system.
Sit in the payment flow out of which repayments can be debited automatically before payments being paid out to the borrower/customer.
Have an already built distribution engine and a set of customers they can sell into already (e.g. a new feature for loan applications could be launched straight into the software UI ideally and online applications received accordingly). I wouldn’t propose lending to customers who had not been a core customer for at least 6-12 months.
Own the future weekly performance data of the customer for the whole of the loan term. If a yoga studio (Mindbody’s customer) or hotel (Mews’ case) is having lower bookings then this would trigger an action through the data and further credit extensions/increases in risk exposure can be prevented.
In terms of the synergies of launching lending as a product extension, take my aforementioned example regarding the stickiness of payments and add lending: Try leaving Mews or Mindbody if all of your customer payments are managed through their systems and you have a good borrowing relationship to help fund capital expenditure/working capital vs. just being a software customer; it’s much harder.
And this further stickiness is not purely driven by high switching costs, it is driven by value, ease and loyalty that this lending program can provide. If your software vendor has also helped you improve your gym studio or hotel by providing credit then that is much more impactful than purely providing software to run your business. Moreover if there are further planned borrowings on the horizon then why go through the effort of building a credit relationship with another vendor when your Mews or Mindbody could do it quicker and with less risk (i.e. you already have a relationship). Arguably you would probably pay a higher price for this versus taking relationship risk with a new lender at a potentially lower price. I haven’t been able to find much available data on the impact of lending on SaaS customer NPS or churn but I have to think there is some impact.
However the real benefit to the software vendor at scale comes in take rate (in essence these are loan fees charged to the bank partner for providing distribution to the end customer). I know Toast uses this model in the US (source), where it has a partner bank that takes the balance sheet risk for the loan book and Toast manages the user flow/customer relationship, with fee returns split between Toast and the partner bank. Square Capital has been able to generate returns of up to 30% APR (according to its website) with losses of 4% on its SMB loan program whilst extending $9 billion in credit to >460K customers to date since launching in 2014. In fact, upon reviewing Square’s Q3’22 earnings data, the lending program generated $57 million in fees/interest in Q3’22 alone, implying an annual run rate revenue of over $200 million for lending (source). Developing an additional synergetic revenue stream of this scale without taking balance sheet risk (Square used the partner bank model too) makes this interesting from an investor standpoint too.
Looking forward, it feels appropriate for larger vertical software companies to layer in additional revenue streams such as lending to maintain revenue growth and increase stickiness, especially given market size is always a risk in vertical solutions. Early adopters such as Toast and Square began using their own capital (source) before moving to a partner bank model likely once they were able to show data on loan program uptake and credit quality. That said, it feels like there is opportunity for software to solve and automate some of the administrative and compliance issues of running a loan program at scale as a vertical software company. What might this look like in terms of functionality:
Cloud content management for loan statements and borrower interaction/notifications
White-label website integration to manage FAQs and other content resources for loan program
White-label website integration to manage user flow of loan applicants through the credit process (KYC, AML, credit background checks via API, document upload etc)
Customer service tooling regarding the loan program - feels like it would need its own customer service flow or at the very least provide knowledge/content to existing customer service reps on the loan program
Regulatory and compliance updates - software tracks and updates around lending regulation and compliance procedures in relevant jurisdiction (US, UK, Germany etc)
Data integrations/export with partner banks (once loan portfolio at scale/risk moved off-balance sheet) to share portfolio performance and analysis
The borrower value chain for lending looks like this:
Summary value chain for lending program
I think the aim would be to build a software solution that makes building, growing and running a loan program as a vertical software company easy and low risk (in essence becoming the lending operating system). The above value chain is more from the borrower’s perspective, however having internal software tools for building a credit/pricing model rules engine (perhaps a good use case for low-code tooling) would be very valuable. Other internal-facing functionality would include compliance/regulatory reporting (depending on geography) and analytics regarding loan book growth, payments/defaults and a 360-degree view of each borrower from a credit perspective (this should integrate with the payments data to understand borrower performance e.g. yoga studio bookings for someone like Mindbody). I do think the vertical software company using this lending software needs to hire credit talent to retain some control over the process and credit quality, especially if the lending begins on balance sheet. That said, the true value of this lending software would be automating manual tasks, monitoring loan book performance and keeping the operating cost for this business line slim whilst creating a good borrower experience through a streamlined online approval and monitoring process.
From a scale perspective, lending starts to become more relevant once a vertical software company is of a certain size (Series C and beyond). A large and diversified enough customer base is required with a clear view to a loan book of over $100 million I think to make the economics and upstart costs work. You also don’t want to lend to every software customer, just the better credits. Providing working capital and short-term improvement loans of up to $250k would mean a portfolio of >500 borrowers (at $100k loan size) to get to some degree of scale (>$50 million annual loan originations). Splitting fees at 50/50 with the partner bank and assuming annual loan returns of 20% less losses of 5% (source) means roughly a 7.5% fee for the vertical software company (equating to $3.75 million of new revenue on $50 million loan originations). Over time lending can become one of many pillars as the company expands its product offering for one specific vertical. The folks at Bessemer Venture Partners have been similarly hot on product expansion’s importance in vertical software and why companies need this “layer cake” strategy of building additional products to cross-sell into existing customers (source). Other approaches to this I have found would be Stripe Capital, which can be used to offer loans to end customers if you are using Stripe’s payments API. They have partnered with a few startups like Glofox (gym software) to offer loans to end customers however the core Stripe Capital business is lending to Stripe customers (e.g. ecommerce sites) not Stripe’s customers’ customers so this feels less competitive.
From an investment perspective, being the lending operating system for vertical software companies feels like a very sticky business with high switching costs. If you think of the success of the ‘business-in-a-box’ platforms like Shopify, then I think of this lending software opportunity as ‘business-line-in-a-box’. Given loan books will grow over time, having a software pricing model based on loan book growth would lead to strong revenue expansion in customers. I don’t think the software should bundle in capital for the customer (more fin than tech), as this should be managed by the customer on balance sheet to begin with and moved off balance sheet over time.
In terms of market size, the end customer market size is much larger in the US than Europe for vertical software: the US vertical software market is valued ~$123 billion (14% CAGR) as of 2022, with SMB solutions holding 36% of this market (i.e. $44 billion: source). The fit for building a diversified loan book is obviously better in the SMB market where the number of customers is much larger and also given these businesses are less served financially than large corporates. In terms of Europe, I can see a future for large vertical software businesses (Europe has good vertical software heritage - source). Market size expands when including European POS players as even these vendors that have a very granular customer base could feasibly launch lending (e.g. Zettle or SumUp) and get to a lot of scale like Square did in the US. Overall, I think this is something most vertical software companies will seriously explore once they pass the Series C stage, and being the ‘business-line-in-a-box’ software for enabling successful lending programs seems like a good idea in the venture space today.
