The Scientific Method for Banks

November 12, 2015 Dallas Wells


I spoke with a banker earlier this week who expressed some frustration with the rest of his team. He feels that they are unreasonably slow to make decisions, and it ends up delaying important initiatives.

Does this sound familiar to anyone else? I had to reluctantly admit to the guy that this exact issue was probably the biggest reason I decided to jump ship and move outside the industry. I understand the WHY behind this slow process. In most businesses if you are right 95% of the time, you are a runaway success. In banking that same 95% means the bank fails and you’re looking for a job. The low margins and high leverage combine to create an environment where caution is the only right approach, and tiptoeing into uncharted waters will likely earn you a wrist slap from your regulators.

Banks need to find a way to move faster, without compromising their safety. They need to make decisions quickly and then adjust later when those decisions are not perfect. They can learn from startups.

Today’s startups have learned to use the scientific method to make business decisions. They break big projects and decisions into small parts, each of which individually do not have the potential to be fatal. They then form a reasonable hypothesis and test it. If the hypothesis is proven correct, you move on to the next piece. If the hypothesis is wrong, you scrap it and try again. They have learned to fail fast on pieces that will not cause too much pain, and to build things piece by piece.
Think about software. In the “old days” (you know, 6 or 7 years ago), software came out in giant releases. Updates were saved up until they all got released at once in a shiny “version 2.0” that is then sent out into the wild to be used by paying customers. When bugs inevitably showed up, they were tangled up in the giant release, and bug fixes wouldn’t be ready until version 2.1 is released.

windows-versionsNow, in many shops (including ours), software updates get released as soon as they are ready. They are broken into small pieces, built, tested, and released. If there are bugs, or the new piece doesn’t play nice with existing code, it is easy to find the culprit. It can either be fixed or taken down and rebuilt. Once it has proven to be stable and useful, the next step can be built and released, and your product evolves into something provably stable, useful, and valuable over time.

So, how does this concept translate to banking? Banks actually have a very fertile testing ground. They have thousands of customers doing thousands of transactions every month. Instead of guessing about product design or pricing, why not break it into pieces and test it? Fail fast on painless small tests instead of failing on a massive campaign that we spent piles of money and man hours building.
Wondering how commercial borrowers will react to aggressive pricing on floating rate loans to offset a building interest rate risk problem? Test the rates for a month with a select few lenders and see what happens. Learn from the feedback and adjust the hypothesis.

Wish you knew how depositors will react when rates start rising? Test a CD special in one market for a couple of weeks, and see how many of those “surge deposits” parked in transaction accounts move to the higher rate. Then try the same special in another market with a new account designed to stem that flow and see if it works. All of these tests are FAR cheaper than getting it wrong across the entire balance sheet once things start moving.

A/B TestingThe big takeaway here is that it’s okay to be cautious. I think we all know the implications for banks that abandon that caution, as those problems always come home to roost eventually. But, maybe we should rethink what being cautious really means. It doesn’t have to mean sitting in hours of committee meetings trying to design the perfect plan. After all, as Mike Tyson said, “everyone has a plan until they get punched in the mouth.”

Instead, the truly safe approach is to embrace some entrepreneurial spirit. Hire smart people and then give them the room to take some small, incremental chances. If you fail, do it small and do it fast. When you don’t fail, you have something you can build on, and over time you will leave your slow moving peers in the dust behind you.

The post The Scientific Method for Banks appeared first on PrecisionLender.


Previous Article
Social Listening for Bankers – Your Brand is Not Your Own Anymore [Podcast]
Social Listening for Bankers – Your Brand is Not Your Own Anymore [Podcast]

A couple of weeks ago we released part one of our discussion with Charlie Johnson from Flagstar Bank in Mic...

Next Article
IACPM Annual Fall Conference, November 2015
IACPM Annual Fall Conference, November 2015

Last week Jenny Bo, our Director of Client Development, Carl Ryden, our CEO and Co-founder and Ken Garcia, ...