How I See This Market Now

A while back I sat next to a woman on a flight who worked at one of the twenty largest banks in the country. She asked why anyone would build a company to sell software to a bank with four branches.

So I got to explain it. Community institutions lend to the local contractor nobody else will touch, because they know him and they know his work. They hold the deposits in the town they sit in, so the money stays near the people who built the place.

She had spent her whole career in banking and nobody had ever put it to her that way.

If somebody inside the industry has not heard it, the founders raising money on a story about serving banks and credit unions have not heard it either. They run a sales process built for a buyer they never actually studied.

I did the same thing for a long time. Twenty years selling software into these institutions, and for most of it my forecast was a big list, a lot of activity, and a number at the end of the quarter that was part math and part hope. When a deal went quiet I called it timing and moved on.

Timing was the right word. I had no idea what set it.

The whole market fits on one screen

Pull FDIC BankFind and the NCUA 5300 call report on the same day and you are holding every institution in the country. I refresh mine quarterly. There are 4,233 banks and 4,336 credit unions. 8,569 in total.

At the end of 2016 there were 11,854. So 3,285 are gone in under ten years, a drop of 27.7 percent. Since 2019 this country has chartered 70 new banks while 1,255 institutions stopped existing.

Every founder I show that to reads it as a dying market. It isn't. The deposits did not vanish, they moved into whoever was left standing. Only 32 banks hold more than $100 billion in assets, and 3,804 of the 4,233 banks still carry the FDIC's community bank flag. Thousands of institutions, most of them small, all of them still buying software.

The counting is the part that matters. Every one of those 8,569 has a name, a charter number, an asset size and a call report anybody can read.

Almost nobody knows their real number

Ask a fintech doing $8 million how many institutions can actually buy their product and you get a range. Usually it is the full 8,569, or some round fraction of it that came out of a board deck two years ago.

The real number is smaller and it is knowable. What the product does narrows it. Who inside the institution owns that problem narrows it again. Then asset size, because a product that assumes a compliance officer and a marketing person on staff is not sellable to a credit union with eleven employees, no matter how good it is. Score all 8,569 against that and you stop guessing.

I do that before I look at anybody's pipeline. Doing it in the other order wastes a month.

What has to be true before they can buy

Something structural has to be true before an institution can buy. A contract has to expire. Or an examiner writes a finding that frees up money. Sometimes it is just a board meeting that has not happened yet. Which condition governs you depends on what you sell, and if you don't know which one it is, your forecast is full of deals that were never going to close inside the window you promised.

Core, core adjacent and digital banking answers to the incumbent core contract. Five to seven years, with the real opening about eighteen to twenty four months before term end. Those deals look healthy right up until the day they go silent.

Risk and compliance runs on exam findings. That gate opens rather than closes. Money shows up suddenly with a deadline attached, which is why those teams have quarters nobody in the building can explain.

Lending waits on the origination system incumbent, then on volume and the rate environment. Deals stall in integration scoping nobody priced into the cycle.

Payments and treasury overlays die in finance. Somebody runs the interchange math, works out that the new revenue does not cover what the institution gives up on its existing processor agreement, and kills it in a meeting the rep was never in. What the rep hears back is that it is not a priority right now. He writes down "no budget," forecasts it again next quarter, and never finds out that the number was the whole problem.

Data and marketing tools have the lowest gate in the market, which is the budget calendar. Easiest thing here to sell. Also the easiest to push a year. Lands in January, dead by October.

Four more apply no matter what you sell. Vendor management review runs on their calendar and it is the most common place a good deal quietly disappears. There is a dollar figure above which your deal waits for a board that meets monthly or quarterly, and most reps never ask what that figure is. Miss the budget window and you wait a year. And a $300 million institution might have two people capable of running an implementation, already booked through spring.

What I had wrong about core conversions

I used to treat a core conversion as a reason to walk away from an account. Plenty of people still do.

That is backwards. An institution evaluates a new core twelve to twenty four months before it converts, and that stretch is the only window in seven years when they are genuinely in market. Walking away means leaving right when they start looking. Give those accounts a stage of their own, out of the quarterly forecast and out of the graveyard, and work backwards from the evaluation date rather than the conversion date.

One honest note, since I see this claimed a lot. Which core a bank runs is findable. When the current contract expires is not public and you cannot buy it anywhere. You get that by asking, and almost nobody asks.

The dead quarter

Somewhere after a good demo and before a decision, your champion is inside the institution writing a business case you will never see.

Most teams do not measure that stretch. They should. How much pipeline is sitting in it, how long it has been there, and what share of it ever comes out the far side. The answer is always worse than anybody guessed, and it explains more slipped quarters than any rep behavior I have ever seen.

The methodology nobody checks

This is the one that annoys me most.

Companies spend real money training a team on Sandler or Challenger or Gap Selling or MEDDIC, and two quarters later nobody can say whether the reps use it or whether it did anything.

The reason is boring. Whatever the method asks a rep to do, the CRM usually has nowhere to record it. If the method wants the economic buyer documented and there is no field for the economic buyer, that step is invisible and unmanaged, so it fades. The people who did try it drifted off, because nobody ever showed them their own win rate went up while they were doing it.

So I built a check for it. Tell me which method your team was trained on. I will list what that method requires a rep to actually do, then look at whether your CRM records any of it. If it cannot, that is finding number one and I will tell you which fields to add. After that it gets easy to show each rep what the method did to their own numbers, which is the only version of this argument that has ever changed anybody's behavior.