Spreadsheets are not the enemy. We say this on every kickoff call. The enemy is hidden spreadsheets: the ones that quietly hold a piece of the truth, are owned by one person, and would take a week to reconstruct if that person left.
The job of a custom dashboard is not to wipe out spreadsheets. It is to retire the dangerous ones, and let the useful ones live in peace. Here is how we stage that.
Why not just delete them
Because the spreadsheet is the muscle memory. The person who maintains it has built a workflow around it across years. The formulas encode small decisions nobody wrote down. Wipe the file on day one and you wipe the institutional knowledge that made the dashboard buildable in the first place.
A clean kill takes weeks. A safe retirement takes days. We pick retirement.
The four phase plan
Phase 1: Inventory
Before we touch a thing, we list every spreadsheet that touches the workflow we are dashboarding. Filename, owner, cadence, what it is for, what would break if it disappeared on Monday.
On a recent build, the answer was "four". By the end of the inventory, it was nineteen. Most teams underestimate by a factor of three.
Phase 2: Classify
Every spreadsheet falls into one of four buckets.
- System of record.The data lives nowhere else. Migrate to the dashboard's warehouse before anything else. The spreadsheet stays read-only for thirty days as a fallback.
- Sync target. The data lives in a real system, the spreadsheet is a manual copy. The dashboard replaces it on day one. Original spreadsheet is archived, not deleted.
- Working surface.Someone's scratch pad for personal use. Leave it alone. These do not need to die.
- Report. Output that someone else reads. The dashboard absorbs it, but the spreadsheet keeps generating for sixty days in parallel so we can prove the numbers match.
Phase 3: Parallel run
For at least one reporting cycle, the dashboard and the spreadsheet run side by side. The owner of the spreadsheet keeps maintaining it. We add a column on the dashboard called vs sheet that shows the delta. When the delta sits at zero for two cycles, the spreadsheet retires.
The owner has to trust the numbers before the spreadsheet can die. No exceptions.
Phase 4: The funeral
A short ritual matters. We move the spreadsheet into an Archived folder, lock it read-only, name it FILENAME_retired_YYYY_MM_DD, and send a one-line note to the team: this sheet is now retired, the dashboard is the source of truth, the archive stays for two years.
That note is what makes the kill safe. Anyone who reaches for the old file finds it, sees the note, knows where to go next.
The spreadsheets we never kill
Some spreadsheets earn a stay of execution. Recognise them and stop fighting them.
- Personal calc-pads someone uses to think before they enter a number into a real system.
- One-off models for a deal or a hire. They will be abandoned within a quarter on their own.
- Anything finance built that has to match a tax report. Their format is the format, not yours.
What it feels like, week three
Three weeks after a dashboard goes live, the operations lead usually says some version of: I have not opened the old sheet in five days. That is the moment we ask permission to start the funeral.
Until then, the spreadsheet stays. Patient retirement is the whole pattern. Kill it too fast and the team will rebuild it somewhere we cannot see.
