Skip to content
EncryptTechBD
Home Industries Our Work Blog Pricing About Contact Discuss Your Project
বাংলা (BN)
Business Software 6 min read

Data migration quality checklist: a working template

A working data migration checklist with owners, evidence, reconciliation tests and sign-off gates for scope, mapping, rehearsal and cutover.

E
EncryptTechBD · 0 comments
Desktop computer and monitor on a wooden desk
Photo: Anita Peeples / StockSnap

A data migration is finished when the business can use the new system, not when the import job reports success. Quality has to be designed into the work: scope agreed, source data profiled, mappings reviewed, rehearsals reconciled and a rollback tested. This checklist is a working document. Copy it into the project workspace, add names and dates, and refuse sign-off at any gate where the evidence is missing.

How to use this checklist

Each gate has an accountable owner and a required evidence item. Record the evidence link beside the item and the date it was reviewed. A gate closes only when the owner and the business process owner have both signed. If an item is not applicable, write down why rather than leaving it blank. Complete Gates 0 to 2 before the first rehearsal, re-run Gates 1 to 4 at every rehearsal, and review Gate 5 weekly for the first month after go-live.

Gate 0: Scope, ownership and data inventory

  • Systems and entities in scope. Owner: migration lead. Evidence: signed scope list with exclusions.
  • Business acceptance criteria. Owner: process owner. Evidence: test scenarios tied to real work.
  • Data owners by domain. Owner: project sponsor. Evidence: RACI covering customers, finance, inventory and documents.
  • Retention and legal hold. Owner: legal or compliance. Evidence: retention schedule and hold register.
  • Risk register and rollback owner. Owner: project manager. Evidence: approved register with named escalation contacts.

Gate 1: Profiling and quality baseline

Profile the source before writing transformation logic. The baseline becomes the comparison point for every later rehearsal.

  • Row counts per table or file. Owner: data engineer. Evidence: timestamped count report.
  • Null and blank rates for required fields. Owner: data analyst. Evidence: field profile with sample records.
  • Duplicates by business key. Owner: data analyst. Evidence: duplicate report and agreed merge rule.
  • Orphan records and broken references. Owner: data engineer. Evidence: referential integrity report.
  • Invalid codes, dates and currencies. Owner: business analyst. Evidence: exception list by source system.
  • Bangla and mixed-script text. Owner: data engineer. Evidence: encoding test with real sample documents.
  • Sensitive data map. Owner: security lead. Evidence: classification for personal, financial and credential data.

Gate 2: Mapping and transformation rules

  • Field mapping from source to target. Owner: business analyst. Evidence: versioned mapping workbook approved by both system owners.
  • Code and value mappings. Owner: business analyst. Evidence: lookup table with unmatched values listed.
  • Defaults and derived fields. Owner: data engineer. Evidence: rule description and unit tests.
  • Date, time zone and currency rules. Owner: data engineer. Evidence: worked examples across boundary cases.
  • Tax and rounding behaviour. Owner: finance representative. Evidence: reconciliation of sample invoices.
  • Rejected record handling. Owner: migration lead. Evidence: queue, owner and correction route for every rejection.
  • Version-controlled scripts. Owner: data engineer. Evidence: repository commit, peer review and release tag.

Gate 3: Rehearsal and reconciliation

Run the full migration more than once, including the cutover steps. A rehearsal that skips the freeze or the rollback is not a rehearsal.

  • Row counts by table. Owner: QA lead. Evidence: source-to-target comparison with variance explained.
  • Control totals. Owner: finance or operations. Evidence: sums, counts and hashes for balances, orders and documents.
  • Sample comparison by stratum. Owner: business tester. Evidence: high-value, high-risk and random samples checked field by field.
  • Business rule tests. Owner: process owner. Evidence: signed test results for pricing, credit, tax and approvals.
  • Exception log. Owner: QA lead. Evidence: every rejected record with an owner and resolution.
  • Volume and timing test. Owner: data engineer. Evidence: load timings against the cutover window.
  • Rollback rehearsal. Owner: cutover manager. Evidence: timed restore and verification of the old system.

Gate 4: Access, privacy and audit

  • Least-privilege access for the migration team. Owner: security lead. Evidence: access matrix and approval records.
  • Masking in non-production environments. Owner: data engineer. Evidence: masked sample and re-identification test.
  • Encryption in transit and at rest. Owner: infrastructure lead. Evidence: configuration evidence for every data store.
  • Vendor and contractor access. Owner: security lead. Evidence: named accounts, time limits and session logs.
  • Audit trail of changes. Owner: security lead. Evidence: retained logs for migration scripts, imports and manual corrections.
  • Deletion after migration. Owner: privacy lead. Evidence: certificate of deletion for temporary copies and extracts.

Gate 5: Cutover, rollback and post-go-live

  • Business freeze and final delta plan. Owner: cutover manager. Evidence: freeze notice and delta capture log.
  • Restore-tested backup. Owner: infrastructure lead. Evidence: restore record with date and verification.
  • Rollback trigger and decision time. Owner: project sponsor. Evidence: written trigger, owner and deadline.
  • Runbook with named steps. Owner: cutover manager. Evidence: minute-by-minute plan with contacts.
  • User and support communication. Owner: change lead. Evidence: messages, briefing notes and support rota.
  • Go or no-go meeting. Owner: project sponsor. Evidence: signed decision with open risks listed.
  • Day-one and week-one verification. Owner: process owner. Evidence: reconciliation results and issue log.
  • Legacy system retirement. Owner: infrastructure lead. Evidence: read-only or decommission date and archive location.

Sign-off record

  • Gate 0: migration lead, process owner, date, evidence location.
  • Gate 1: data analyst, data owner, date, evidence location.
  • Gate 2: business analyst, finance representative, date, evidence location.
  • Gate 3: QA lead, process owner, date, evidence location.
  • Gate 4: security lead, privacy lead, date, evidence location.
  • Gate 5: cutover manager, project sponsor, date, evidence location.

Failure patterns this checklist is designed to catch

Count-only reconciliation hides value errors, especially when the number of rows matches but amounts or statuses do not. Testing with cleaned sample data hides the exceptions that will appear in production. A migration without a rejected-record queue leaves staff to fix problems in spreadsheets after go-live. Security and retention are often added at the end, when temporary copies have already spread across laptops. And a rollback that has never been timed is a hope, not a plan.

Migrations that span several systems need a data engineer and a business owner working from the same evidence. Our custom software team plans migrations with reconciliation reports and a tested cutover. The pricing page sets out how engagements are structured; the AI-ready data foundations guide explains the downstream effects.

Sources

Services that deliver this

Want this implemented properly?

Pricing →

Every engagement starts with a written scope. If a requirement is outside it, we quote it separately before starting.

Responses (0)

No comments yet. Add the first one below.

Leave a response

Your email address will not be published. Required fields are marked *.

Keep reading

More field notes