# The custom tool we shouldn't have built

Three weeks of build. Six months later, two formulas in Google Sheets would have done it.

We built a custom reporting tool last year. Took about three weeks.

Six months later I realised Google Sheets with two formulas would have done 90% of what it does. The other 10% we never actually use.

Now we have a rule: use what exists until the friction is measurable. Not "I wish it did X" — that's a preference. "We lose hours every week because it can't do X" — that's friction.

Still learning to follow the rule. Builders like to build.