The end of the consolidation spreadsheet (and what replaces it)
Every company has a spreadsheet someone rebuilds weekly by stitching exports together. It is expensive, fragile and always out of date. What has to exist to retire it without losing control.
It almost always exists and almost never appears on the org chart: the spreadsheet someone assembles on Friday from four exports, fixing campaign names by hand and pasting into a summary tab. It is the most used information system in most mid-sized companies — and the most fragile.
Why it costs more than it looks
- Recurring assembly cost. Two to six hours a week from someone hired to do something else.
- A human single point of failure. That person goes on holiday and the week has no numbers.
- Silent errors. A paste one row off, a forgotten filter, a month counted twice — none of it warns you.
- Version conflicts. Two copies circulating by email and the meeting becomes a debate about which is right.
- Freezing. The moment it is sent, it is already old.
The consolidation spreadsheet is not disorganisation. It is the patch that exists because the sources meet nowhere. Until they meet, it comes back.
What has to exist to retire it
- 01Sources ingested automatically, with no manual export. While someone downloads a CSV, the work only moved.
- 02People resolved as people. That is what replaces the hand-made email lookup.
- 03Criteria written once. What counts as a sale, what is a renewal, which currency — decided in the data, not in a cell formula.
- 04A page that reads the base live. Whoever opens it sees the current number; nobody sends attachments.
- 05A connected assistant for anything off-pattern. Without it, every new question rebuilds the spreadsheet.
Steps 4 and 5 travel together, which is why the replacement usually fails when only one is done. A dashboard without an assistant pushes exceptions back into Excel; an assistant without a fixed page removes the ritual the company uses to align.
A migration that works
| Week | What to do | Success criterion |
|---|---|---|
| 1 | Connect the sources and reproduce ONE number from the spreadsheet | It matches, or the difference is explained |
| 2 | Reproduce the rest of the summary and run in parallel | Two consecutive weeks with no unexplained gap |
| 3 | Publish the live page and use it in the meeting | Nobody opened the spreadsheet during the meeting |
| 4 | Retire the spreadsheet and keep the history | The person who assembled it has another task |
Running in parallel for two weeks is not excess caution: it is what turns "I trust the tool" into "I checked the tool".
When to keep the spreadsheet
There are legitimate cases: simulation, budgeting, hypothetical scenarios. Spreadsheets are great at inventing numbers that do not exist yet. What they should stop doing is transporting numbers that already exist elsewhere.