No more open-ended man-day invoices.Buy outcomes, not man-days.We offer Implement at a fixed price based on the validated scope.

MindDX Blog

Data Cleansing Before Moving to ERP: A 5-Step Plan

TopicsERPDigital Transformation

MindDX cover: a typographic graphic with the article title and four of its sections, from «Why Before Migration?» to «Conclusion».

The most expensive misconception in ERP migrations is this: "The consultant will clean the data while migrating it anyway." Data migrated without cleansing produces the old errors faster in the new system: risk limits split across duplicate customer accounts, inventory inflated by non-standard stock codes, automations that do not work because of missing fields. In AI-supported systems this becomes even more critical: a model cannot exceed the quality of the data it is fed. Gartner, in its ERP predictions report, predicts that 70% of organisations will lack AI-ready ERP data by 2027. The good news: data cleansing is not a talent but a discipline — and it can be managed with a five-step plan.

Why Before Migration?

Cleansing during migration is the most expensive cleansing: it is done under the pressure of the project schedule, in the rush of field mapping, with "we’ll fix it later" decisions — and "later" never comes. Cleansing done before migration, on the other hand, is both cheap and speeds up the migration itself: what is to be migrated becomes clear, mapping becomes simpler, test data becomes reliable.

A 5-Step Data Cleansing Plan

1. Inventory and ownership

Which master data areas do you have (customers, products, suppliers, bills of materials, price lists), and who owns each one? A data area with no owner is an area nobody cleans. The first output: a table of area × owner × number of records.

2. Define the standards

Coding rules (stock code structure, customer numbering), mandatory fields and format standards (phone, tax ID, unit) are put in writing. Cleansing done without defined standards is cleansing according to personal preference — the next person dirties the data again.

3. Cleansing duplicate and dead records

Three customer account records for the same customer under three different spellings; stock codes with no movement for five years; open records of staff who have left. The rule: duplicates are merged, dead records are archived — not deleted (they are needed for historical reports), but they are not migrated to the new system as active.

4. Enrich and validate

Missing mandatory fields (tax office, payment terms, unit weight) are completed; critical fields are validated at their source. The output of this step is measured: the "mandatory field completion rate" is tied to a pre-migration target.

5. Set up a mechanism to sustain it

Cleansing is not a one-off: the authority to create new records is narrowed, entry rules are defined in the system, and periodic (monthly/quarterly) data quality checks are assigned to the owners. Cleansing without a mechanism is back to square one within six months.

3 Common Mistakes

  1. The "the consultant will clean it" assumption: the consultant maps and migrates; only your own team knows which of your records is correct. Cleansing cannot be bought in from outside; it can only be managed.
  2. Migrating everything: every dead record migrated "in case it is needed" pollutes the new system’s reports from day one. The archive stays in the old system or in the data warehouse.
  3. Leaving cleansing to the end of the project: if test scenarios are run with dirty data, the test itself becomes unreliable — cleansing must be finished before testing.

Conclusion

Data cleansing is the highest-return item in the ERP budget: at relatively low cost, it reduces both the migration risk and the cost of errors during operation. To see where to start, the 2-minute ERP Risk Score shows your risk in the data dimension and the Free Digital Maturity Scan your overall picture.

Related articles: The X-Ray of AI Promises · How Is the Cost of an ERP Project Calculated?

To make sure you do not miss new analyses, subscribe to the MindDX-Digital Excellence newsletter.

Talk to MindDXAsk the AssistantAll articles on MindDX Blog

Start your project with a measurement, not an estimate.

Talk to MindDX