A Dynamics 365 Business Central implementation takes 4 to 7 months — 16 to 30 weeks. A multi-country or heavily customised Finance & Operations programme takes 9 to 18 months. The spread inside each band is decided almost entirely by two things on the client side: how clean the source data is, and how fast decisions get made. Neither is controlled by the partner, which is why a timeline quoted without seeing your data is a guess.
Microsoft Dynamics 365 30-Week Milestone Governance Timeline
Phase 1: Diagnostic & Blueprint
Deliverable: Solution Blueprint & Fit/Gap MatrixPhase 2: Design & Prototyping (CRP)
Deliverable: Iterative Conference Room PilotsPhase 3: Development & Integrations
Deliverable: Custom Extensions, APIs & Migration ETLPhase 4: UAT & End-User Training
Deliverable: UAT Sign-off & Mock Cutover RunPhase 5: Cutover, Go-Live & Hypercare
Deliverable: Live Production & 24/7 HypercareThe five phases, and who owns each
| Phase | Typical window | Client owner | What "done" means |
|---|---|---|---|
| 1 · Diagnostic & blueprint | Weeks 1–4 | Finance lead | Chart of accounts and dimension structure signed off in writing. |
| 2 · Design & conference room pilots | Weeks 5–10 | Process owners per area | Each process demonstrated against real data, not sample data. |
| 3 · Development & migration build | Weeks 11–20 | IT lead | Sandbox trial balance reconciles to source, account by account. |
| 4 · UAT & training | Weeks 21–26 | Named testers per process | A full period closed in the sandbox with matching numbers. |
| 5 · Cutover & hypercare | Weeks 27–30 | Project sponsor | First month-end closes clean in production. |
The client owner column is the part most plans omit. A phase without a named person on your side is a phase where the partner makes assumptions, and every one of those assumptions surfaces in UAT when changing it is expensive.
Where timelines slip
- Discovery signed off without the chart of accounts. The single most common cause. Everything downstream is configured against a structure that then changes.
- Migration scoped as an extract. Moving records is fast. Making them reconcile to the trial balance you are leaving is the work, and it is routinely under-scoped by half.
- UAT run as a demo. Walking through screens is not testing. A full period close against real volumes is, and it finds things nothing else will.
When a date is at risk, the pressure lands on the reconciliation phase because it looks like checking, not building. Compressing it is how a variance becomes a permanent feature of your accounts.
What to protect and what to trade
If the date cannot move, something else must. In order of what costs least to trade: scope of non-standard customisation, number of integrations in phase one, historical data depth, and number of entities going live together.
What should never be traded is the reconciliation gate and the sandbox period close. Those two are what tell you the date was real.
After go-live
Hypercare runs until the first month-end closes cleanly in production. Go-live weekend proves the data loaded; the first close proves it was right. Treating handover as the end of the weekend is how teams discover problems in month two with the partner already demobilised.