Cloud Migration Checklist Excel: Pre, During and Post Steps

12 min read · 2026-11-18

A good cloud migration checklist excel file has one row per task. Each row has a phase, an owner, a due date, a status, a dependency and proof that the task is done. Next to it, keep a second sheet for your risk register. Every risk there gets a likelihood, an impact, a score, an owner and a trigger that tells you when to act.

Most checklists online stop at a list of steps. This guide gives you the full working file. You get the columns to copy into Excel or Google Sheets and filled example rows you can reuse as they are. You also get the tasks for before, during and after cutover, a scored risk register and a phased timeline to show your CIO, your client or your steering committee.

The roadmap at a glance

Goal: Move a first wave of applications to the cloud with tracked tasks, scored risks and a tested rollback plan. Duration: 5 to 6 months

  1. Strategy and business case (Weeks 1-2)

    Agree on why you are moving and what success looks like before anyone touches a server.

    • Write down the business objectives: cost, resilience, speed of delivery, or the end of a data center contract.
    • Name the sponsor, the migration lead and an owner for each application.
    • Define the success metrics you will check after migration, such as uptime, response time and monthly spend.
    • Set the decision rules for budget and scope changes.

    Milestone: Business case and success metrics signed off by the sponsor.

  2. Discovery and inventory (Weeks 3-5)

    Know exactly what you are moving and what depends on what.

    • List every application, database, server and integration in one inventory sheet.
    • Map dependencies between workloads, including batch jobs and file transfers.
    • Record data volumes, data sensitivity and where the data must legally stay.
    • Keep discovery running instead of doing it once, so new dependencies show up when they appear.

    Milestone: Inventory complete, with an owner and a dependency map for every workload.

  3. Migration approach and target design (Weeks 6-8)

    Pick a strategy for each application and design where each one will land.

    • Choose a strategy for each application. AWS Prescriptive Guidance lists seven, known as the 7 Rs: retire, retain, rehost, relocate, repurchase, replatform, and refactor or rearchitect.
    • Design the target architecture: accounts, networks, identity and connectivity back to on premises.
    • Group applications into migration waves based on dependencies and risk.
    • Estimate the cloud cost of each wave.

    Milestone: Wave plan approved, with a strategy and target design for every application in wave one.

  4. Security, compliance and risk register (Weeks 7-8 (in parallel))

    Find what could block or slow the move while there is still time to fix it.

    • Build the risk register and score every risk.
    • Check regulatory needs: HIPAA, PCI DSS, data residency and any contract clauses.
    • Define access rules, encryption, logging and backup policies for the target.
    • Agree on the acceptable downtime window for each application with its business owner.

    Milestone: Risk register reviewed. No high score risk is left without a mitigation.

  5. Pilot migration and testing (Weeks 9-11)

    Prove the method on a low risk workload before you scale it.

    • Migrate one or two low risk applications end to end.
    • Run functional, performance and security tests against agreed thresholds.
    • Rehearse the rollback plan and time it.
    • Update the runbook with what went wrong.

    Milestone: Pilot live, tests passed and rollback rehearsed within the agreed window.

  6. Cutover waves (Weeks 12-19)

    Move the rest of wave one in controlled batches.

    • Take and verify backups before each cutover.
    • Freeze changes on the applications being moved.
    • Run the cutover runbook step by step, with a named person on each step.
    • Hold a go or no go check at a set time. If the criteria are not met, rollback is the default.

    Milestone: All wave one applications running in the cloud, source systems in read only or switched off.

  7. Stabilize and optimize (Weeks 20-23)

    Make sure the move delivered what the business case promised.

    • Monitor performance and errors against the success metrics.
    • Rightsize instances and remove unused resources.
    • Compare actual spend with the estimate, line by line.
    • Decommission old servers and close the contracts.

    Milestone: Success metrics reported to the sponsor and old infrastructure retired.

Cloud migration checklist in Excel: the columns to set up

A checklist only works if anyone can open it and see what is late and who owns it. Set up one sheet called Tasks with the columns listed below, in this order.

Add a data validation list on the Status column so people cannot type their own values. Then use conditional formatting to turn a row red when the due date has passed and the status is not done. A filter on Owner gives each person their own view without a separate file.

Most templates skip the Proof column. It is also the column your auditor, your client or your CISO will ask about first. "Done" without evidence is only an opinion.

  • Task ID: a short code such as DISC-04, so tasks can point to each other.
  • Phase: strategy, discovery, design, security, pilot, cutover or optimize.
  • Application or workload: the system the task applies to.
  • Task: one action, written with a verb ("Export firewall rules").
  • Owner: one person, not a team.
  • Due date: a real date, not "week 3".
  • Status: not started, in progress, blocked, done.
  • Depends on: the Task ID that must be finished first.
  • Proof of completion: a link, a ticket number or a test report.
  • Notes: decisions and blockers.

A filled Tasks sheet: five example rows

Here is what five rows look like once they are filled in. The dates assume a first wave that starts in January 2027. Copy the pattern, then swap in your own systems, people and dates.

Look at what makes these rows usable. Each owner is a role you can turn into one name. Each task can only be checked in one way. Each dependency points to a real Task ID. When CUT-01 slips, you can see at once that CUT-02 slips with it.

  • DISC-04. Phase: discovery. Workload: billing app. Task: map every dependency of the billing app, including nightly batch jobs and file transfers. Owner: finance systems application owner. Due: 12 February 2027. Status: in progress. Depends on: DISC-01 (inventory sheet complete). Proof: dependency diagram attached to the billing row of the inventory.
  • SEC-03. Phase: security. Workload: customer database. Task: confirm the target region meets the data residency rule for customer records. Owner: security architect. Due: 26 February 2027. Status: done. Depends on: DES-02 (target accounts created). Proof: change ticket number and a screenshot of the region setting.
  • PIL-02. Phase: pilot. Workload: intranet portal. Task: rehearse the rollback of the intranet portal and record how long it takes. Owner: migration lead. Due: 19 March 2027. Status: not started. Depends on: PIL-01 (intranet portal migrated to the pilot environment). Proof: rollback test report with start time, end time and result.
  • CUT-02. Phase: cutover. Workload: order database. Task: take a full backup of the order database and test a restore on the day. Owner: database administrator. Due: 9 April 2027. Status: not started. Depends on: CUT-01 (change freeze active on the order system). Proof: backup job ID and restore test log.
  • OPT-05. Phase: optimize. Workload: all wave one applications. Task: compare the first month of actual spend with the wave one estimate. Owner: finance business partner for IT. Due: 28 May 2027. Status: not started. Depends on: CUT-07 (last wave one cutover signed off). Proof: Costs sheet updated, with the gap shown for each line.

Before, during and after cutover: the checklist items

The tasks change a lot depending on where you are in the move. Splitting them helps you run the right meeting at the right time.

Before migration, the checklist is about knowledge and decisions. During migration, it is about control. After migration, it is about proof.

  • Before: inventory and dependency map complete.
  • Before: strategy chosen for each application.
  • Before: target accounts, network and identity ready.
  • Before: backup and disaster recovery approach agreed.
  • Before: downtime window agreed with each business owner.
  • Before: rollback plan written and rehearsed in the pilot.
  • During: backups taken and restore tested on the day.
  • During: change freeze active.
  • During: each runbook step timed and signed off by its owner.
  • During: go or no go decision logged with the time and the name of the person who made it.
  • During: users and support teams told when the switch happens.
  • After: performance and uptime compared with the baseline.
  • After: security logging and alerts confirmed working.
  • After: costs tracked against the estimate.
  • After: old systems switched off and data deleted where required.
  • After: lessons learned added to the runbook before the next wave.

Cloud migration risk assessment template

Create a second sheet called Risks. Use these columns: Risk ID, Category, Description, Likelihood (1 to 5), Impact (1 to 5), Score, Mitigation, Owner, Trigger and Status. The Trigger is the early sign that a risk is turning into a problem. It keeps the register useful for the whole project, not only at kickoff.

A practical way to sort migration risks is to use five groups: technical blockers, regulatory requirements, performance and availability, links between migrated and non migrated workloads, and security. Here is one example row for each, written the way you would enter it.

  • Technical blocker: the billing database uses stored procedures that block a move to a managed database service. Mitigation: rehost that database first and plan the refactor later. Trigger: the compatibility scan flags unsupported features.
  • Regulatory: patient records will be stored by a cloud provider without a signed HIPAA Business Associate Agreement. Mitigation: sign the agreement before any data moves. Trigger: a data transfer is scheduled and no signed agreement is on file.
  • Performance and availability: an order system responds more slowly once its database sits in another region. Mitigation: move the app and its database in the same wave and region. Trigger: response time in the pilot goes above the agreed threshold.
  • Links between workloads: a nightly batch job on premises still reads files from a server that has moved. Mitigation: map the job in discovery and set up secure connectivity before cutover. Trigger: the job fails or runs late after a wave.
  • Security: new cloud storage is created with public access by mistake. Mitigation: apply a policy that blocks public access by default. Trigger: the configuration scan reports a public resource.

How to score risks: likelihood times impact

Score = likelihood × impact, both on a scale of 1 to 5. Write down what each number means once, at the top of the sheet, so two people give the same score to the same risk. The scale below is a good starting point.

Set three bands. A score of 15 or more is red: it needs a mitigation and an owner before the next phase starts. A score from 8 to 14 is amber and is reviewed every week. A score below 8 is green and is reviewed at each phase gate. These bands are a working convention, not a standard, so agree on them with your sponsor.

Review the register at every weekly meeting. Risks move. A low score in discovery can jump once the pilot shows real performance numbers.

  • Likelihood 1: very unlikely. 3: could happen in this project. 5: expected unless we act.
  • Impact 1: minor delay, no effect on users. 3: one wave slips or one service degrades. 5: data loss, a compliance breach or a business stop.

Compliance, data residency and costs

Compliance decides what can move, where it can go and in what order. Check three things early: which data is regulated (health data under HIPAA, card data under PCI DSS), which country or region it must stay in, and which contracts limit where your data may be processed. Each one becomes a task with an owner and proof, such as a signed agreement or a region setting confirmed in the target account.

Costs need the same discipline as tasks. Before migration, estimate the monthly cost of each wave. During migration, watch for double running costs, since old and new systems often run side by side for a while. After migration, compare actual spend with the estimate every month and rightsize anything that runs far below its capacity. Add a Costs sheet with three columns for each wave: estimate, actual and gap.

When a spreadsheet is enough and when to add a visual roadmap

A spreadsheet is enough when one team runs one wave, the task list fits on a few screens and everyone reading it already knows the project. Filters, validation and conditional formatting cover the daily work well.

Keep the spreadsheet as the working file and the record of proof. Add a visual roadmap on top of it once you have several waves, several teams, or a steering committee to report to. That roadmap shows the phases on a timeline, the milestone that closes each one, and the main links between them. Update both at the same weekly meeting so they never tell different stories.

The spreadsheet starts to fall short in three places.

  • Timeline view. Rows sorted by date do not show how long each phase lasts, where phases overlap or how far the cutover sits from today. You can build a Gantt chart in Excel, but it is easy to break every time a date changes.
  • Dependencies. The Depends on column records one link at a time. It does not show a chain. When DISC-04 slips, nothing on the sheet shows that the pilot, the first cutover and the cost review all move with it.
  • Presenting to leadership. A sponsor or a client will not scroll through 200 rows. They want the phases, the milestones and the decisions they need to make, on one page.

Presenting the plan and its risks to leadership

Leaders do not want your 200 row spreadsheet. They want to know when, how risky and how much. Turn the checklist into a one page roadmap with the phases on a timeline, the milestone at the end of each one and the current wave highlighted.

Next to it, show the top five risks by score, each with its owner and mitigation on one line. End with the decisions you need from them, such as approving a downtime window or signing a contract. Send the visual before the meeting and keep the spreadsheet as the backup for detailed questions.

Common mistakes to avoid

  • Running discovery only once at the start. Fix it by refreshing the inventory and the dependency map before every wave.
  • Giving a task to a team instead of a person. Fix it by putting one name in the Owner column.
  • Marking tasks done with no evidence. Fix it by making the Proof column mandatory before the status can change.
  • Writing a rollback plan nobody has tested. Fix it by rehearsing it in the pilot and recording how long it took.
  • Scoring risks without a shared scale. Fix it by defining each number from 1 to 5 at the top of the risk sheet.
  • Treating cost as a finance topic after go live. Fix it by tracking estimate against actual from the first wave.

Frequently asked questions

What should a cloud migration checklist include?

It should cover strategy, discovery, the approach for each application, target design, security and compliance, pilot, cutover and optimization. Each task needs an owner, a due date, a status, a dependency and proof of completion. Keep risks on a separate sheet so they get their own review.

How do you do a cloud migration risk assessment?

List risks by category, score each one on likelihood and impact from 1 to 5, and multiply the two. Give every high score a mitigation, an owner and a trigger. Repeat the assessment as discovery and testing bring new facts, not just once at the start.

What are the 7 Rs of cloud migration?

AWS Prescriptive Guidance lists seven strategies: retire, retain, rehost, relocate, repurchase, replatform, and refactor or rearchitect. You pick one for each application based on its value, its technical fit and the effort you can afford. Record the choice and the reason in your inventory sheet.

Is Excel good enough to manage a cloud migration?

For a first wave with one owner per task, a well built spreadsheet works. It needs data validation on statuses, conditional formatting for late tasks and a separate risk sheet. Once you have to show the timeline, the dependency chains or the progress to leadership or a client, add a visual roadmap on top.

Generate this roadmap with AI