The spreadsheet that won’t die
Every company has one. It refuses to die because it works — and that deserves respect before anyone tries to replace it.
Every company has one spreadsheet that is quietly running the company. It has a tab called “final_NEW_v2”. One person understands it completely, and that person is on vacation.
It has outlived at least one serious attempt to replace it. That is not stubbornness — it is a recommendation. The spreadsheet was shaped, over years of tiny corrections, into an exact mould of the process the business actually runs. Most software never gets that close.
So before replacing it, it’s worth knowing what would actually justify that.
When to replace it
A few signs, and they compound. Several people need it open at the same time, so the file gets emailed around and edited in parallel copies that disagree by Thursday. An error in it costs money silently — a wrong price that gets honoured, a wrong quantity that gets ordered. Somebody official asks who changed what, and when, and the honest answer is a shrug. Or the spreadsheet has quietly become infrastructure: quoting, payroll, inventory — load-bearing, and owned by no one.
Any one of these is a warning. Two together is a decision waiting to be made.
When to leave it alone
One person, one purpose, occasional use, and mistakes that become visible quickly. That is not a risk. That is a tool, doing its job, for free. Swapping working tools for newer ones is fashion, and it costs software money forever.
The migration trap
The instinct, when the spreadsheet finally is replaced, is to rebuild it faithfully. Every tab, every quirk, every workaround — including the bugs, which get ported with the same care as the features. What you end up with is an exact copy of the spreadsheet, at software prices, on software’s schedule.
A migration done right is a deciding exercise, not a copying one. What job is this thing actually doing? Which of its steps are real business rules, and which are scar tissue from the spreadsheet’s own limitations? What would the process look like if it were designed today, by the people who run it, instead of accreted one formula at a time since whenever it started?
Answer those on paper first. Then build the answer.
The spreadsheet earned its place. Replace it when the job outgrew it, not when something shinier came along.
Working on something this applies to? Tell us about it — thirty minutes, no pitch.