Implementation Timeline Roadmap by Project Type: Cloud Migration Example

7 min read ยท 2026-11-16

An implementation timeline roadmap puts the phases of a rollout on a calendar, with an owner and a finish line for each one. The phases are not the same for every project. A software rollout, a cloud migration and a process change each need a different timeline, and the cloud migration is the one that most often gets planned as a single big date.

This guide shows how the timeline changes by project type. It then gives a complete cloud migration implementation timeline roadmap: a 24-week plan built in migration waves, with named owners, cutover windows, rollback points and the switch-off of old servers. It is written for IT leads, project managers and consultants who have to hand this plan to leadership or a client.

The example scenario: a mid-size company moves its on-premises file servers, an internal web application and a reporting database to a public cloud. It keeps the same applications and changes as little as possible during the move. The 24 weeks (about 6 months) are a planning default for this scenario, not a benchmark. Adjust them to the size of your inventory.

The roadmap at a glance

Goal: Move the agreed workloads to the cloud in controlled waves, with no unplanned downtime, and shut down the old servers. Duration: 24 weeks

  1. Discovery and inventory (Weeks 1-3)

    Know exactly what you are moving before you pick a single date.

    • List every server, application, database and file share, with its business owner.
    • Map which systems talk to which, including scheduled jobs and reports.
    • Note the downtime each system can accept and the busy periods to avoid, such as month-end close.
    • Choose an approach per workload: move as is, adjust slightly, rebuild, or retire.

    Milestone: Inventory and dependency map signed off by IT leadership.

  2. Wave planning and business case (Weeks 4-5)

    Group the workloads into waves that can move together.

    • Put systems that depend on each other in the same wave.
    • Start with low-risk workloads, keep the most critical ones for later waves.
    • Give each wave a target weekend, a cutover owner and a rollback deadline.
    • Agree the budget owner and how cloud costs will be tracked after the move.

    Milestone: Wave plan approved, with a date for every wave.

  3. Landing zone and security (Weeks 6-9)

    Build the cloud foundation before anything moves.

    • Set up accounts, networking, identity and access rules.
    • Connect the cloud network to the office network.
    • Set up backups, monitoring and logging.
    • Have security review the setup against company policy.

    Milestone: Landing zone passed by the security review.

  4. Pilot wave (Weeks 10-12)

    Move one low-risk workload end to end to test the method.

    • Migrate an internal file share or a test environment.
    • Time each step of the cutover and write it into a runbook.
    • Test the rollback for real, not on paper.
    • Fix what the pilot reveals before wave 1.

    Milestone: Pilot workload running in the cloud, runbook updated.

  5. Migration waves 1 to 3 (Weeks 13-20)

    Repeat the tested method, one wave at a time.

    • Wave 1 (weeks 13 to 14): remaining file shares.
    • Wave 2 (weeks 15 to 17): the internal web application and its integrations.
    • Wave 3 (weeks 18 to 20): the reporting database, outside the month-end period.
    • For each wave: freeze changes on the old system, migrate, test with business users, then confirm or roll back before the deadline.

    Milestone: Each wave confirmed in writing by its business owner.

  6. Optimization (Weeks 21-22)

    Tune what now runs in the cloud.

    • Right-size machines that are larger than needed.
    • Check that backups and alerts work on every migrated system.
    • Review the first cloud bills against the budget.

    Milestone: Cost and performance review accepted.

  7. Decommissioning and closure (Weeks 23-24)

    Switch off the old world so you stop paying for two.

    • Take a final archive of the old servers.
    • Shut them down and update the asset register.
    • Hand day-to-day ownership to the operations team.
    • Hold a review and record what to change for the next migration.

    Milestone: Old servers retired and closure report signed.

How the implementation timeline changes by project type

The same word, "implementation", covers very different projects. Before you draw the timeline, decide which type you have. The type tells you where the time goes and what the risky milestones are.

For a generic phase-by-phase plan with owners and gates, the guide on building a roadmap covers the base method. This page focuses on the shape that a migration needs.

  • Software or system rollout (CRM, ERP, HR tool): time goes into configuration, data cleanup and user testing. The timeline is linear: build, test, pilot with one team, then roll out by department. The risky milestone is user acceptance, because that is where real users show what the build missed.
  • Cloud or infrastructure migration: time goes into discovery and into the waves themselves. The timeline repeats: the same move, done several times on different workloads. The risky milestones are the cutover weekends, so each one needs its own date, owner and rollback deadline.
  • Process change (new approval flow, new review cycle): there is little to build. Time goes into writing the process, training and getting people to use it. The timeline follows the business calendar: the pilot has to fit one real cycle, such as one month-end or one review period.
  • Compliance or policy implementation: the deadline is fixed from outside, so you plan backwards from it. Time goes into gap analysis, evidence and sign-offs. The risky milestone is the internal audit before the real one, which must leave time to fix findings.
  • Office or site move: the calendar is driven by contracts, suppliers and one move date. Most tasks lead up to a single switch weekend, with a short settling period after it.

Why a cloud migration needs waves, not one go-live date

A single go-live date puts every risk on one weekend. If one system fails, everything waits. Waves split the risk. Each wave is small enough to roll back, and each one improves the runbook for the next.

Waves also make the roadmap easier to read. Leadership sees a series of dated moves with a clear order. Business owners see the exact weekend when their system changes and what they must test on Monday morning. Three rules make waves work.

  • Group by dependency, not by department: an application and the database it reads must move together, or you create slow links between the cloud and the old data center.
  • Order by risk: low-risk workloads go first to test the method. Critical workloads go last, once the runbook is proven.
  • Respect the business calendar: never schedule finance systems near month-end or retail systems in peak season.

Cutover windows and rollback points

Every wave on the roadmap carries three dates, not one.

Write the rollback decision into the plan before the wave starts. Name who decides, what triggers a rollback (for example, failed user tests or missing data) and how long switching back takes. A rollback that has been tested in the pilot is a real option. One that has never been tested is a hope.

  • Change freeze start: no new changes on the old system from this point.
  • Cutover window: the hours when the move happens and users are warned of downtime.
  • Rollback deadline: the last moment you can switch back to the old system safely.

Owners in a cloud migration

A migration crosses teams, so make ownership visible on the timeline.

The business owner is the one most often missing. Without them, nobody confirms that the system really works for its users.

  • Migration lead: owns the overall roadmap and the wave dates.
  • Cloud or infrastructure engineer: owns the landing zone and the technical move.
  • Security lead: approves the landing zone before any data moves.
  • Business owner per application: tests their system after each cutover and confirms it works.
  • Operations team: takes over day-to-day running at closure.

Common mistakes to avoid

  • Starting dates before the inventory is done means you will find hidden dependencies mid-wave, so finish discovery first.
  • Planning one big cutover puts every risk on one weekend, so split the move into waves you can roll back.
  • Skipping the pilot wave hides gaps, because the first real move always reveals steps missing from the runbook.
  • Leaving out a rollback deadline keeps teams fixing on Monday morning while users wait, so set one for every wave.
  • Forgetting decommissioning leaves old servers running and means paying twice, so put the switch-off on the roadmap with a date and an owner.
  • Ignoring the business calendar turns a small issue into a big one, so never schedule a cutover during a busy period.

Frequently asked questions

What is an implementation timeline roadmap?

It is a plan that shows the phases of an implementation on a calendar, with the dates, an owner for each phase and the milestones that prove each phase is finished. It answers both "what is the order?" and "when does each step happen?".

How long does a cloud migration implementation take?

It depends on how many workloads you move, how tangled their dependencies are and how much you change them during the move. Size it from your inventory: count the waves you need, add discovery, the landing zone, the pilot and decommissioning, then add buffer before the critical waves.

What are the phases of a cloud migration roadmap?

A common order is discovery and inventory, wave planning, landing zone and security, a pilot wave, migration waves, optimization, then decommissioning and closure.

How is an implementation timeline different for a software rollout?

A software rollout is mostly linear: configure, test with users, pilot with one team, then roll out by department. A migration repeats the same move in waves, and its riskiest moments are the cutover weekends.

Generate this roadmap with AI