Skip to content
Ixia Consulting

Resource for SAP programme leads

SAP S/4HANA data readiness checklist

What to check before the first mock load: who owns it, what evidence to capture, and what should block sign-off. Read it here, or take the worksheet into your programme.

Free to download. No email form. Published by Ixia Consulting, November 2026.

Before the first mock load
  1. 1. Scope and ownership

  2. 2. Source extracts and profiling

  3. 3. Business Partner and duplicate resolution

  4. 4. Material master checks

  5. 5. Reconciliation

  6. 6. Mock-load sign-off

How to use this checklist

  1. Open the worksheet and fill in the yellow cells for each check in your scope: the data object, a named owner, the pass criteria and where the evidence is kept.
  2. Agree an owner and a pass threshold for every row with the business before the first mock load, and write them down.
  3. Re-run the checks every cycle. Keep the evidence with the cycle scorecard, so sign-off rests on something a business owner can read.

Thresholds are programme-specific

This page deliberately gives no pass percentages. What counts as a pass depends on the object, the programme and the risk the business is willing to carry. Agree each threshold with the data owner and record it before the first mock load.

01

Scope and ownership

Settle what moves, and who answers for each object, before anything is extracted.

Typically owned by

Programme data lead, with a named business owner for each data domain. Scope changes go to the steering committee.

What to check

  • Every in-scope data object is listed with its source, its target object and how it will be loaded.
  • Out-of-scope data is listed too, with the reason. Closed historical transactions need a retention decision, not an assumption that they migrate.
  • Each data domain has a named business owner who can make mapping and survivorship decisions.
  • The rule for changing scope between mock cycles is agreed, along with who can approve a change.
  • The place where cleansing happens is decided. Where possible, fix in the source so the improvement outlives the go-live.

Evidence to capture

  • Object inventory showing what is in scope, what is out, and why.
  • Owner register with a name against every domain, not a department.
  • Decision log for scope, and for the retention or decommissioning of historical data.

Escalate or block sign-off when

  • An in-scope object has no named owner.
  • Scope changed after profiling and the cycle plan was not revisited.
  • Historical data is assumed to migrate with nobody having decided that it should.

Thresholds: programme-specific.

Historical data that should not migrate still needs a plan. See SAP legacy decommissioning.

02

Source extracts and profiling

Measure the source as it is, not as the design workshops believe it to be.

Typically owned by

Migration lead for extracts and profiling. Business data owners for the rules and for cleansing decisions.

What to check

  • Each extract is dated and can be regenerated with the same filters and the same result.
  • Extract totals agree to the source system in the control totals the programme chose.
  • Every in-scope object is profiled before mapping starts: distributions, blanks, outliers and structure.
  • Profiling findings are compared with the assumptions in the design. Where they disagree, someone has ruled.
  • Business rules are written down for what the target process needs, separating technically valid from business valid, each with an owner, a severity and the reason it exists.
  • Cleansing work is sized and assigned, with an owner against each item.

Evidence to capture

  • Extract log: source, date, filters and row counts.
  • Profiling report for each object.
  • Rule register with owner, severity and business reason.
  • Cleansing backlog with an owner and a target cycle for each item.

Escalate or block sign-off when

  • An in-scope object has not been profiled.
  • An extract cannot be reproduced.
  • A profiling finding contradicts the mapping and nobody has ruled on it.
  • A critical rule has no owner.

Thresholds: programme-specific.

Why profiling comes first: data assessments change the scope, not just the schedule.

03

Business Partner and duplicate resolution

Customers and vendors converge in S/4HANA, so every unresolved duplicate becomes a conversion decision.

Typically owned by

Business data owners from finance, procurement and sales, with the Business Partner design owner and the migration data lead.

What to check

  • Duplicate candidates are identified across customers and vendors, including one legal entity held under several numbers.
  • Merge and survivorship rules are agreed with the business before the conversion strategy is fixed. Which value wins is a business decision.
  • Records are merged automatically only where the evidence is strong. The grey zone is queued for a data steward.
  • Inactive records, test accounts and other agreed exclusions are listed rather than silently dropped.
  • Every merge decision keeps lineage back to the source records.

Evidence to capture

  • Duplicate cluster report.
  • Survivorship rule register, with the owner who agreed each rule.
  • Merge decision log with lineage to the source records.
  • Count of unresolved pairs, by cycle.

Escalate or block sign-off when

  • Survivorship rules are not agreed by the time the conversion approach is frozen.
  • Records were merged automatically on weak evidence.
  • Unresolved duplicate pairs are not falling from one cycle to the next.

Thresholds: programme-specific.

Background: what survivorship rules actually decide.

04

Material master checks

A material that is technically valid can still be unusable for the process that has to buy, stock, pick or sell it.

Typically owned by

Material master data owner from supply chain or product data, with the migration data lead.

What to check

  • Base unit of measure is populated and consistent with any alternative units.
  • Material type, group and description are valid and populated for the target.
  • Barcodes (EAN or GTIN) are present where the business uses them, correctly formed, and not duplicated across materials.
  • Organisational extensions are complete for where each material trades: plant, sales organisation, distribution channel and division.
  • Cross-object checks pass. For example, a material extended to a sales area has a delivery plant.
  • Reference and check values that the target validates more strictly than the source did are tested before load.
  • Failures are counted at business grain: one material counted once, however many plant rows it sits on.

Evidence to capture

  • Rule results by object and rule, at business grain.
  • Failure file naming the record, the field, the current value and how it is corrected.
  • Trend per rule across mock cycles.
  • A list of rules that could not run, reported as blocked.

Escalate or block sign-off when

  • A critical rule fails above the threshold the programme set.
  • A rule could not run because its data was missing, and it has been treated as a pass.
  • Failures are counted per row, so one defect looks like many or real defects are hidden.

Thresholds: programme-specific.

More on rule design: what a data quality rule needs to say to be auditable.

05

Reconciliation

Counts reconcile long before values do. Agree what the evidence must prove before the first load.

Typically owned by

Reconciliation lead for the packs. Finance and business owners for balances. A named business owner signs.

What to check

  • The reconciliation approach is agreed before the first mock load, not designed after it.
  • A source baseline is captured at extraction.
  • Completeness, control totals, record-level values and relationships are reconciled between source and target, not only record counts.
  • Where values are deliberately changed, the change matches a documented rule.
  • Every rejected record is listed, explained and dispositioned.
  • Merged and exceptional records are accounted for individually.
  • Every fix is re-reconciled, because a fix that is not re-proved is a newer assumption.

Evidence to capture

  • Reconciliation pack for each cycle: control totals, value exceptions and relationship checks.
  • Rejected-record list with a disposition against each.
  • Named sign-off from an owner who can read the evidence.

Escalate or block sign-off when

  • Reconciliation rests on record counts alone.
  • A rejected record has no disposition.
  • A fix has not been re-reconciled.
  • The person asked to sign cannot read the evidence.

Thresholds: programme-specific.

What the evidence must prove is set out under data migration reconciliation. On the SAP Retail programme in our case study, Ixia delivered the migration workstream and provided reconciliation evidence for each migration.

06

Mock-load sign-off

A load that ends without errors proves the tooling worked. Sign-off is the business saying it received its data.

Typically owned by

Programme data lead runs the review. Named business owners sign for their objects.

What to check

  • Entry criteria for the mock load are agreed, and the status of sections 1 to 5 is known before it starts.
  • The cycle runs at real volume, so timings and load behaviour are measured, not assumed.
  • Data owners review the results object by object.
  • Open issues are listed, each with an owner and a target cycle.
  • The processes that depend on the data run on the migrated data: invoicing, payments, picking and reporting.
  • Issues trend towards zero from one cycle to the next.
  • Accepted risks are recorded with a named owner.

Evidence to capture

  • Cycle scorecard: open issues by severity, and the trend.
  • Signed acceptance for each object.
  • Run timings from the cycle.
  • List of accepted risks, each with an owner.

Escalate or block sign-off when

  • A critical issue is open without an owner.
  • The trend is flat or worsening across consecutive cycles.
  • The cycle did not run at real volume, so its timings cannot be trusted.
  • A risk is accepted without a named owner.

Thresholds: programme-specific.

Why a clean load is not enough: loaded is not migrated.

Illustrative example, fictional data

What a failed check looks like, and what closes it

All names and figures below are invented to show the shape of a failed check. They are not results from any client.

Check
Customer tax or VAT number is unique across active customers.
Result in mock cycle 2 (fictional)
61 of 4,280 active customer records share a tax number with at least one other record.
Business impact
Duplicates would convert into duplicate Business Partners. Credit exposure would be split across numbers, and statements and dunning notices could go to the wrong record.
Owner
Head of credit control (customer data owner), with the migration data lead.
Evidence needed to close
The cluster report listing each pair. A keep or merge decision for every pair, signed by the owner, naming the surviving number. A re-run of the check showing every remaining match explained, for example branches that legitimately share a group tax number. The updated survivorship register.
Effect on sign-off
Blocks Business Partner conversion sign-off until closed, or until the owner formally accepts the residual risk.

Take it into your programme

An Excel workbook with every check on this page ready to track: owner, pass criteria, evidence, status and next action for each, with drop-down statuses, a live summary by area and a worked example.

sap-s4hana-data-readiness-worksheet.xlsx

Sheets in the workbook

  • Start here
  • Checklist
  • Owners and escalation
  • Summary
  • Worked example

The Checklist sheet has a row for each of the 37 checks. Statuses: Not started, In progress, Passed, Failed, Blocked, Accepted risk.

Where Ixia fits

Ixia migration services

Ixia delivers the migration workstream: discovery and profiling, mapping, rehearsed mock cycles, value-level reconciliation and cutover support, with named owners and evidence the business can read.

SAP S/4HANA data migration, data migration reconciliation and migration assurance.

Ixia DQ, the product

Ixia DQ profiles source objects, runs business-defined rules, detects duplicates and keeps the trend for each rule across mock cycles. It does not extract, transform or load, and it does not reconcile source to target.

See Ixia DQ for migration readiness.

Want a second pair of eyes before the first mock load?

Talk to a senior practitioner about your migration readiness, or see how Ixia DQ measures source data against your own rules.