01
Step one: divide the work into phases
Start by splitting the project into a few phases with clear exits: discovery, design, data preparation, configuration and testing, training, cutover and stabilization. Write one sentence describing what must be true for each phase to close. In the distributor example, discovery ends when every order-to-cash and purchasing process is documented and the team agrees which ones change. Avoid creating a phase for every module at first. Fewer, well-defined phases make it easier to see dependencies, such as data cleanup needing to finish before test cycles begin, and they give executives a simple view of progress.
02
Step two: assign an owner to every deliverable
Every deliverable needs one named owner, even when many people contribute. For the distributor, the warehouse manager owns inventory counts and location data, the controller owns the chart of accounts and opening balances, and a sales lead owns customer records. The implementation partner supports but does not own business data. Add a backup for each owner to cover absences. Record owners in the same document as the phases so questions about delays have an obvious first contact. Unclear ownership is a frequent cause of slipped dates, so settle it before work begins.
03
Step three: set milestones and review points
Choose milestones that can be checked, not just felt. Examples for the distributor include cleaned customer and item data approved, two full test cycles passed with signed-off scripts, staff training completed for each role, and a go or no-go meeting held one week before cutover. Place a short review at the end of each phase where owners report open risks and decisions needed. Keep the milestone list to a page. If a date moves, update dependent items the same day, so the plan remains the single reference the whole team trusts.