The first ninety days of a roadmap

A modernization roadmap earns trust or loses it in its first quarter. What to schedule in days 1 through 90, and what to refuse to.

Pixel drawing of a map with a dotted route leading to a flag

Roadmaps do not fail in year two. They fail in the first ninety days, quietly, when the gap between the document and the calendar first shows. Whatever the five-year picture says, the first quarter decides whether anyone will still be following the plan when it matters. Schedule it accordingly.

Days 1 to 30: finish the counting. The inventory and the standing-stock price go from draft to signed. The renewals calendar is assembled and the next four quarters of contract tripwires are known. Every deadline with a date certain (leases, end-of-support clocks, regulatory horizons like the SAP 2027 wall that most of the market is still approaching unmigrated) is on one page. No construction yet. Foundations.

Days 31 to 60: one visible retirement and one visible move. Pick the most defensible decommission on the list (the expensive-and-marginal pile always has a candidate) and turn it off, with the savings figure attached. Move one low-stakes workload end to end through the real pipeline (the easy 20% exists for this). The point is not the workload. The point is that the org watches the machinery work: change process, comms, rollback, the checklist, all exercised at lawn-chair stakes.

Days 61 to 90: name the hard thing and start its clock. Every estate has one item everyone knows is the boss fight: the shared database, the WCF estate, the system whose expert retires next year. It does not get solved this quarter. It gets an owner, a seam analysis, a rehearsal budget and a fortnightly heartbeat, so that month eight finds it half-derisked instead of still politely unmentioned.

And the refusals, which make the schedule possible. Refuse the tool procurement in quarter one; platforms bought before the inventory settles are shelfware with a kickoff deck. Refuse the big-bang candidate, whatever executive loves it; the estate earns big moves by landing small ones. Refuse steady-state metrics theater; the quarter reports exactly three numbers: things turned off, things moved, dollars per year removed from the standing stock.

Ninety days later the roadmap is either a document the org quotes or a document the org survived. The difference was never the analysis. It was whether the first quarter kept its own small promises, in public, on dates. Plans borrow trust. Delivery pays it back. Start repaying in week six.