The model I built that nobody used
A perfectly correct tool that failed because I solved the wrong problem.
I once built a fairly careful spreadsheet model to help screen incoming project files. It worked. It was checked. As far as I know, nobody but me ever opened it.
The failure was not technical. It was that I had built a tool for the problem I found interesting rather than the problem the desk actually had. The interesting problem was ranking incoming files by how much attention they were likely to need. The real problem was that files arrive incomplete, and the first thing anyone needs is a fast, boring answer to “is this submission even complete enough to review?”
A checklist would have been more useful than a model. It would also have taken an afternoon.
Three things I took from it:
- Watch what people do before building anything. Not what they say the process is, but what actually happens in the first five minutes with a new file.
- The most valuable tool is usually the least clever one. If it is not obvious within thirty seconds how to use it, it will not be used, whatever it does.
- A tool that requires a new habit will fail. The successful version has to fit inside a step that already exists.
I am keeping this note visible because I can feel myself about to make the same mistake with the timeline dataset, building the interesting analysis rather than the thing someone would actually pick up.