What can be verified, and what must be asked.
Every company we work with has its own definition of an ideal customer. We sort that definition in two. The traits that can be established from public evidence become a research standard, measured on every company in your market, every month. The ones that can’t become the questions worth asking in a first conversation.
This page is the standard we built for one real market, so you can see exactly what survives that sorting and what doesn’t. The company doing the selling is TieOut, a demo company we invented, because we won’t put a client’s market on display. The market, the companies and the records are all real.
One rule sorts a definition in two.
A trait belongs in the research standard only if public evidence can settle it, without anyone having to ask. Not answerable in principle. Answerable in practice, across a whole market, in the same month. It does not have to come back on every company, and the share each one was actually found on is printed beside it below.
That rule is stricter than it sounds, which is why a research standard is always shorter than the definition it came from. What it turns away isn’t lost. Those traits become questions your team asks in discovery, where one person can ask one company directly and have an answer in a minute. The work of research is the half that would take a year to ask.
First, who is in your market at all.
Before anything is measured, a few questions settle which companies count as being in your market at all. For this example there are three. They run before any other evidence is read, so every company shown here passes all three, which makes them a description of the boundary rather than a finding about the market.
US insured depository
It is a US bank whose deposits are federally insured. Insured banks file public financial reports every quarter, which is where most of the evidence here comes from.
FDIC BankFind, the public register of insured institutions.
Assets between $300M and $10B
Big enough to run an operations team, small enough that it buys software rather than building its own.
The quarterly call report every insured bank files.
At least fifty employees
Big enough to have people whose whole job is the back office, which is who this software is sold to.
Full-time equivalent headcount, on the same call report.
Then, six things worth measuring.
These six survived the rule for this example, whose customers are US community banks. Yours would be different, because a standard is built out of your definition of an ideal customer rather than out of a template. Each is scored on every company in the market, every month, and the figure beside it is the share where public evidence confirmed it.
- Confirmed on the record
- Couldn’t be answered either way
Efficiency ratio elevated
Its costs are high against its income, and rising. Banks report this number every quarter and watch it closely, so a bad one already has the board’s attention.
Published quarterly in the call report, with three years of history behind it.
Headcount absorption strain
It has added people far faster than it has added business. That is what taking on work you cannot yet absorb looks like from the outside.
Headcount and total assets, both on the call report, compared across three years.
Conversion or operations hiring
It is advertising for the roles that a system conversion needs: loan operations, deposit operations, reconciliation. Banks staffing up for that work usually show it before they talk about it.
Job postings the bank has published itself.
Unabsorbed acquisition workload
It has bought another bank and has not finished merging it in. Not that a deal happened, but that the work is visibly still open.
FDIC structure change records for the deal, and the bank’s own website for whether the seams are still showing.
Customer-facing digital layer mid-change
Customers of an acquired bank still log in through a separate page. It is the most visible sign that a merger is unfinished, and anyone can see it.
The bank’s live website. No access required, and no guessing.
Competing platform installed
It is already running rival software of this kind, so there is something in the way. Rare, and the trait the public record answers least often.
Public technology traces on the bank’s own estate, which do not always carry.
13 criteria, read across 1,921 companies. As of 21 September 2026.
And four ways to leave your market.
A company can stop belonging in your market after it has been added, so the entry questions are scored in reverse as well. One that grows past the top of the size band, or slips below the bottom of it, is caught on the next monthly run instead of sitting there unchallenged for a year. Nothing in this example has left yet.
- Below $300M in assets
- Above $10B in assets
- Credit union
- Non-US institution
What the rule turns away, and where it goes.
Building this one, we tried eight traits that didn’t survive. Each came off for one of three reasons, and each was settled by a number rather than by an argument. None of them was wasted. A trait that can’t be measured across a whole market is still worth knowing about one company, so it moves from the research standard to the discovery conversation, where asking is cheap.
Nobody publishes it.
What we wanted to know was real and the public record simply didn’t carry it. Wanting an answer badly enough doesn’t make it findable, and one that comes back unknown can’t carry weight in a score. These make excellent discovery questions, because a prospect can answer them in one sentence.
Legacy core on a sunset path
Banks do not publish which core banking system they run, and vendors do not announce a retirement date on a schedule anyone outside can read. The research could almost never answer it.
Core platform identified
Same problem, one step earlier. A trait that comes back unknown more often than not cannot carry weight in a score, however much it would matter if you knew it.
Core platform comparison across the merger
It asked whether the two merging banks ran the same system, so it inherited the blind spot above twice over. It was replaced by looking at the seams the merger left on the website instead.
It was true of almost everyone.
A trait that lights up on half of your market, or lights up for two opposite reasons at once, doesn’t tell anyone where to start. Volume isn’t the same thing as signal. These tend to be worth describing in your positioning rather than scoring in your research.
Branch network changing
It fired just as readily on a bank opening new branches as on one folding branches together. More than half the companies lit up, and for opposite reasons, so it separated nobody.
Multi-bank holding company
True of some good prospects and some poor ones in roughly equal measure. A fact about how a company is built is not automatically a fact about whether it will buy.
Operating leverage gap
It was measuring, less directly, the same pressure the efficiency ratio already measures. The ratio is published quarterly and is harder to argue with, so the overlap was cut.
It answered a different question.
The trait was measuring something real and useful, just not whether a company is a fit. These didn’t get deleted so much as relocated, and they’re still recorded on every company for the team that needs them.
Recent acquisition
A deal that closed and was cleanly absorbed leaves no work behind. The date of an acquisition turned out to say much less than the state of it, so the trait was rewritten as unabsorbed workload.
Operations leadership named
A good question about who to open with, and the wrong question for whether a company is a fit. It was moved out of the standard and is still recorded on every company.
What this standard cannot do.
It can’t tell you that a company will buy. It tells you that a company looks like the ones that already did, on facts anyone can check, and it keeps that separate from whether anything has just changed there. Those are two different questions, so they stay two different scores, fit and timing. A perfect fit sitting still isn’t the same account as a scramble you can’t serve.
It also can’t see inside a company. Everything here stops at the edge of the public record, and where the record says nothing we leave the gap showing rather than filling it in. Three traits above came back unanswered on a small number of companies, sixty-one answers in all, and they’re printed that way.
Where your own standard would start.
Your market plan is the first pass at this, written for your market rather than for somebody else’s. It names which market is worth owning first, what sets you apart inside it, and which parts of your own definition public evidence could carry today. We write it from your website before we have spoken, so it says where it’s assuming rather than knowing.