The software nobody uses: spotting it before you build it

A project can fail in two ways. The first is technical: it does not work, it is slow, it breaks. That shows up immediately and gets fixed.

The second is worse, because it does not show: it works beautifully and nobody opens it. Six months of work, everything delivered, and three weeks later people are back on the spreadsheet.

The three signs, visible beforehand

1. The people who will use it have never been in the room

The project was decided by management, designed with whoever proposed it, approved by whoever pays. The people who will work in it eight hours a day see it on handover day.

This is the strongest sign of all. If no one who actually does that job was present during the prototype phase, the system will answer an imagined process instead of the real one — and the gap between the two is where all the exceptions live.

2. It adds steps instead of removing them

Do the count, literally: how many clicks it takes today to do the thing, and how many it will take afterwards. If the number goes up, however tidier the system is, people will go back to the old method the moment nobody is watching.

A system that asks for the same data twice is already dead on the day it is switched on.

3. The benefit belongs to somebody else

The warehouse hand records movements so the owner has the numbers. The engineer fills in the report so accounts can invoice. In both cases one person does the work and another gets the benefit.

It only works if the person doing the work gets something: fewer phone calls, the report filled in with three taps instead of by hand, the client record already there. If they get nothing, they will do it for as long as somebody checks.

The test that costs one week

Before building, take the person who will use the tool and have them do the job on the prototype, for one real day, with real data. Not a demonstration: the job.

What comes out of that day is worth more than three analysis meetings, because the meetings collect opinions about something imagined, while the test day collects facts about something that exists.

The question to ask at the end

“If this system disappeared tomorrow, what would you do?”

If the answer is “go back to how we did it before, and it would not change much”, you have your answer before spending anything.