Planning a reorg without spreadsheets
The spreadsheet is a fine tool for a roster. It's a terrible tool for a decision. Reorg planning is a decision process — and the failure modes are predictable.
Where the spreadsheet breaks
Version drift. The moment you model a second option, you have two files. By the time leadership has weighed in, you have nine, and the differences between them live in people's memories.
Silent math errors. Sum ranges don't follow people when rows move between teams. The savings number in the deck is whatever the formula happened to catch the night it was exported.
No diff. A spreadsheet shows state, not change. "What exactly is different between option B and today?" is the question every stakeholder asks, and the answer is manual archaeology.
Everyone sees everything. Sharing the file means sharing every salary in it. So instead, nobody shares the file, and the people who need context plan blind.
The model that works: baseline + overlays
Keep one canonical picture of the org — the baseline — and express each option as a set of changes against it: reassignments, pay changes, cuts, additions. This is how version control transformed software, and it maps perfectly onto org planning:
- Every option is always in sync with reality, because reality is the shared ancestor.
- The diff is the plan. "Option B: 4 moves, 2 cuts, 1 backfill, −$180k/yr" is generated, not maintained.
- Reverting a bad idea is deleting an overlay, not reconstructing a file.
- Access control can be real: comp visibility enforced by the server, with everyone else planning in percentages.
A concrete workflow
- Import the roster once. CSV in, org chart out. Fix reporting lines by dragging.
- One scenario per option. Name them what leadership calls them ("Flatten platform", "Cut 8%, protect on-call").
- Let the math argue. The compare table shows cost, headcount, and savings side by side; the change matrix shows exactly who is affected in which option.
- Pressure-test with skills coverage — the step spreadsheets can't do at all. More on that in Finding your bus factor.
- Save a named version before each review, so "the plan the CFO approved" is a restore point, not a filename.
The spreadsheet's job was never the planning — it was just the only place the data lived. Give the data a real home and the planning gets an order of magnitude calmer.