When an Off-the-Shelf CRM Runs Out of Room

By the time someone asks me to build a custom CRM, they have usually decided already. The conversation they want is about features. The conversation worth having is whether the existing tool actually failed, or whether nobody ever finished setting it up.

Those two situations feel identical from the inside and have wildly different price tags.

The spreadsheet tells you the truth

Here is the diagnostic that works. Find the spreadsheet.

In almost every business that genuinely needs custom software, there is a spreadsheet holding the real state of things. Somebody maintains it. Meetings run off it. When the spreadsheet and the CRM disagree, people believe the spreadsheet.

That spreadsheet is not a workaround. It is the actual system of record, and its existence is the clearest evidence you will get that the tool you are paying for has stopped modelling how the business works. Everything typed into the CRM after the fact is data entry for its own sake.

If there is no such spreadsheet, and the complaint is that the CRM is clunky or that reports take a while, you probably have a configuration problem and a much cheaper afternoon ahead of you.

What standard tools genuinely cannot do

Modern platforms are far more flexible than most teams exploit. The ceilings that are real tend to be these:

• A pipeline that is not a pipeline. If your work moves in loops, splits into parallel tracks, or reverses, a stage-based model will fight you forever.
• Relationships that are not one-to-many. A contact belonging to several organisations in different capacities breaks most standard schemas immediately.
• Rules that depend on other systems. If eligibility depends on live data from your operational system, a CRM that cannot reach it will always be out of date.
• Regulatory shape. Retention rules, audit trails and residency requirements that a platform simply does not support.

Notice how few of these are about features. They are about shape. Feature gaps get filled by configuration and integration. Shape mismatches do not, and that is where custom becomes the honest answer.

The trap of automating your habits

Custom software is very good at encoding exactly how you work today. That is the entire point, and it is also the risk.

Some of how you work is genuinely advantage: a qualification process refined over a decade, an approach to accounts nobody else takes. Encoding that is worth real money.

Some of it is habit. It is done that way because the last system required it, or because someone set it up in 2018 and left. Encoding that costs money to build and then costs more to unpick later, and it is invisible while you are specifying because everyone in the room considers it normal.

Which is which is worth an uncomfortable afternoon before anything gets built.

Decide reporting first

The most common expensive mistake in these projects is specifying data entry thoroughly and treating reporting as a later concern.

Then the system goes live, leadership asks how conversion varies by source and quarter, and the answer is that the data model cannot say, because source was captured as free text and the timestamps that would answer it were never stored.

Work backwards. Decide the questions you need answered, then design the model that can answer them, then design the screens that collect what the model needs. Doing it in the other order produces a system people diligently fill in and nobody can learn anything from.

Own the data

Whatever you build, the database has to be yours, in your infrastructure, with exports that work and are tested. Not because a vendor is likely to behave badly, but because the cost of any future CRM change is almost entirely the cost of getting the data out cleanly.

A system you can leave is worth more than a system you cannot, even if you never leave it.

The honest recommendation

Configure the standard tool properly first. Give someone two weeks and a mandate. A surprising share of these projects end there, and that is a good outcome.

If you do that and the spreadsheet is still the system of record, the ceiling is real and building is justified. That is worth scoping properly rather than quoting from a wish list, which is what an MVP Roadmap is for.