Many Fintech firms selling to banks have unrealistic expectations. Bankers are born skeptics, and banks don’t change easily. Selling to banks is hard!
When it comes to Fintech firms, especially the newer ones, banks all seem to come from Missouri. “Don’t just talk about it, show me!” So that’s what a Fintech needs to be able to do.
Small Fintech Convinces Large Bank – How?
One of my clients is an early stage startup called Amberoon, founded by CEO Shirish Netke. Amberoon’s flagship product is an AML monitoring tool called Lucre (“dirty money” … get it?). Lucre is a predictive model that uses historical transaction data to risk-rate new transactions for possible money laundering risk.
When I first met Shirish, I had a question for him. The conversation went something like this”:
Graham: Shirish, your product sounds very good. But how do you convince a large bank that your tiny company has the solution to their massive AML monitoring problem?
Shirish: That’s simple – we offer to prove it to them
Graham: But how can you prove something as abstract as a predictive AML modeling solution?
Shirish: Easy – we do a Proof of Concept
Graham: Isn’t that really just a demo? How does that give a bank a compelling business case to buy?
Shirish: We take a large amount of real production data and run it through our model. We show the level of false positives and false negatives with our solution. Then we compare with the actual results with their existing tools and processes. The bank knows the cost of false positives, and can do the math themselves.
Example: Improving AML Monitoring
Let me explain why Lucre matters. Most AML (Anti-Money Laundering) solutions are rules-based. Models have to be tuned very conservatively to avoid missing true money laundering transactions. This results in huge numbers of “false positives”. These are transactions tagged for possible money laundering, that turn out to be legitimate transactions. False positives cost banks millions in operational expense. They are handled manually and so are also error-prone. No matter how conservative the computer monitoring, it is still possible for money laundering activity to be missed. This can be extremely expensive. The highest (and only ten figure) fine was nearly $2 billion for HSBC in 2012, but nine figure fines happen pretty much every year.
Lucre applies historical data analytics, predictive modeling, and intuitive graphic risk representations to transactions. This greatly reduces false positives while further reducing the risk of a false negative. This is a double win, and clearly quantifiable. But just because Amberoon says it works, why would a bank take it on trust?
Proving a Concept
To be clear, what I mean by a Proof of Concept (PoC) in this article is what its name suggests. A Fintech has proposed an interesting conceptual solution. But it needs to be demonstrated sufficiently robustly to offer proof “beyond reasonable doubt”. This is a little different from the traditional internal PoC. But often, what is called a PoC is actually a demonstration or exploration, rather than a proof of anything.
When selling a solution, a successful PoC will have at least the following four elements.
- Sufficient data volumefor a statistically significant demonstration of the solution (and no more). For Lucre this means selecting a long enough time period to represent all meaningful situations. Depending on the size of the bank, that may range from a month or two to as much as year of data.
- Clearly defined comparison metrics: This seems obvious, but isn’t always in practice. The proposed solution is intended to improve some aspect of the bank’s business. There must be a way to measure that improvement in order for a PoC. In Lucre’s case, banks will understand two key metrics. Firstly the level (percentage) of false positives, which drives operational cost. Secondly, the percentage of false negatives (ideally zero). These are actual suspicious transactions that should be reported. Success for Lucre requires improvements in both metrics.
- Reliable baselinefor each metric. A successful PoC is dependent on having a solid baseline against which to compare results. The bank and the Fintech will both need to understand, and be confident in, that baseline. The best metrics are routinely reported to regulatory agencies. Failing that, they should be operational metrics for bank executive management. If metrics have to be derived that aren’t normally used, then this should be done collaboratively.
- Comparison detailthat allows analysis of what drove the results. This is important for a number of reasons. Firstly, there may be a need for tuning of the solution, based on an understanding of where it isn’t realizing the expected result. Banks will generally be patient (up to a point) with this need. Before implementing the solution, the bank needs to understand what is causing the improvement. They also will need to analyze for any risks that might result from the change in process.
The PoC needs to be managed like any project, with pre-defined resources, timeline and budget. The bank will usually pay for the PoC on a time and materials basis. They will already have agreed conceptually that the solution makes sense. They won't agree to the PoC unless they are reasonably sure it will make the desired improvement.
Again, this is unlike a traditional PoC, which is typically a stripped down solution. A traditional PoC is designed to determine whether a concept is worth building out in full. In the way we’re using the term here, a production system is being used. Likely it will be cloud-hosted. It will have full functionality and UI available to the testing bank. With that in mind, a successful PoC becomes very compelling and will more often than not lead to a sale.
Proving the Fintech’s Value Proposition
A bank needs to be sold on the solidity of the business case. They need to be sure there is no unacceptable risk introduced. Sometimes they need a way to prove a solution to their regulators. They need to know the solution will work in a highly integrated production environment.
The burden of proof is on the Fintech. If the bank doesn’t know you, then you have to prove yourself. Getting in the door in the first place is tough. But it only gets harder. You need to know the bank, its business, its problems, and its decision-making approaches.
In the end it will come down to having a clear and provable value proposition that allows the bank to put a business case together. Banks are driven by bottom-line impacts. These may result from improved use of capital, increase revenues, decreased expenses, or quantifiable risk reductions. Fintech firms need to be able to prove their financial value proposition. A Proof of Concept can be a great way to win the business.
Graham, a 30 year banking veteran, runs BankTech Consulting. He is an expert in commercial banking, and provides strategic insight and internal business cases to banks. He works as a fractional Customer Success Executive to Fintech firms, facilitating their partnership with banks.