Buying guide
When to replace the spreadsheet with custom software
Five signs a business process has outgrown a spreadsheet, and the honest test for whether custom software is worth building or a bad idea.
Kaci Cox
This is a published outline, not a finished post. The argument and structure are here so you can see what is coming. It will be written properly rather than padded out to hit a word count.
What this post will cover
When a spreadsheet is genuinely the right tool, and the five signals that it has stopped being one.
Outline
In defence of the spreadsheet
- It is flexible, universally understood, and free. Most processes should stay in one.
- Custom software is the expensive answer and should go last.
The five signals
- Three or more people edit it and version conflicts are routine.
- Someone has become the full-time keeper of the file.
- The same data is retyped into a second system.
- Errors have started costing real money, not just time.
- It cannot answer a question the business needs answered weekly.
The cheaper things to try first
- A shared database tool before a custom build.
- Workflow automation connecting existing tools.
- Fixing the process rather than the tooling — sometimes the spreadsheet is a symptom.
The honest test
- Hours per week times loaded hourly cost times 52, against build plus three years of hosting.
- If it does not clear comfortably, do not build.
What a build actually involves
- Discovery, fixed scope, weekly demos, handoff with runbook and code ownership.
What does not fit
- Replacing accounting systems, anything with certification requirements, projects with no internal owner.
Notes before writing
- The credibility of this post comes from arguing against the sale for most readers.