ERP data cleansing checklist: what to clean and what to keep before migration

An ERP data cleansing checklist should separate master-data cleanup (duplicates, inactive vendors, item codes) from transaction-history deletion. Clean and merge master data, but keep or archive closed invoices, journals and attachments, because that connected history is what auditors, lenders and potential data licensing buyers value.

What should an ERP data cleansing project leave alone?

A cleansing project should clean master data and leave transaction history intact. Duplicate vendors, dead customer records and inconsistent item codes are worth fixing; closed-year invoices, journal entries, purchase orders and the audit trail behind them are not worth deleting, because that history is the part of an ERP that shows how the business actually ran.

Most cleansing scopes are written as "inactive for N years, flag for deletion." That one rule quietly sweeps up both kinds of records. The checklist below separates them so the migration still gets a tidy chart of vendors and customers while the client keeps years of connected operating history. If the client is a US company with 50+ full-time employees at peak (contractors excluded), that history may also be what makes a SourceX data licensing introduction possible later.

Master data versus transaction history: where is the line?

Master data describes who and what the business deals with. Transaction history records what happened. Cleanse the first; archive the second.

Record groupExamplesDefault action in a cleansing project
Master dataVendor, customer, item, employee-of-record, bank and tax fieldsDeduplicate, standardize, merge, deactivate
Open transactionsUnpaid invoices, open POs, unapplied cashMigrate in full
Closed transactionsPaid invoices, receipts, shipped orders, closed projects, posted journalsKeep or archive; do not purge on an inactivity rule
Attachments and notesScanned bills, contract PDFs, approval comments, memo fieldsKeep with the transaction they explain
System logsApproval workflows, user activity, change historyKeep a copy before cutover if the new system will not carry it

The pre-purge checklist

Run these checks before anyone approves a deletion list. Each item is a question the controller or fractional CFO can answer from the ERP without opening a single record's contents.

Scope and ownership

  • The cleansing scope document says which tables are cleaned and which are only migrated.
  • A named owner (controller, CFO or IT lead) signs off on every deletion rule.
  • The earliest transaction date in each module is written down, with the module name beside it.
  • Nobody has scheduled deletion of closed-period data to shorten the migration.

Vendor and customer master cleanup

  • Duplicate vendors are merged, not deleted, so historical invoices still point to one vendor ID.
  • Inactive customers are flagged inactive rather than removed, so closed orders and credit memos keep their parent.
  • A crosswalk of old and new IDs is saved before any merge.
  • Free-text fields on master records are left alone unless the cleanup rule names them.

Transaction history and attachments

  • Posted journals, subledger detail and closed projects are exported or archived before the old system is switched off.
  • Attachments are listed by location and volume, even if they move later.
  • Approval trails and exception notes are included in the archive plan.
  • Anything under a retention hold from an auditor, lender or tax position is marked and excluded from deletion.

Metadata worth capturing

  • Systems feeding the ERP (CRM, payroll, expense tool, ticketing) are listed, with the years each has been connected.
  • The retention setting in each system is recorded.
  • Someone confirms who can still run a full export from the legacy system after cutover.

What do the checklist results mean?

Count the unchecked boxes and act on the pattern, not just the total.

ResultWhat it meansNext action
Scope and ownership gapsDeletion rules may be running without an accountable ownerPause the purge until an owner signs the rule list
Merge-versus-delete gapsHistorical documents may lose their parent recordRe-scope as merge and deactivate, then rebuild the ID crosswalk
Archive gapsClosed-year history may exist only in a system that is about to retireSchedule a full export before cutover and keep it
All checked, 5+ years of history across several systemsThe client may be worth a preliminary fit screenRun the company fit checker with the owner's permission

How does a clean master file connect to a data licensing conversation?

It does not change the client's rights, but it often reveals how long and how connected the records are. A controller who has just mapped the vendor and customer masters can say how many years the ERP covers and which other systems feed it, which are the first two questions in a fit screen.

The data asset register template turns those answers into a metadata-only sheet: system, owner, years held and retention, never record content. If the register shows many systems with long histories, the owner can decide whether to explore a license. Partners never export, upload or describe confidential records, and any redaction or de-identification rules are agreed with the company before work begins.

What to say when the cleansing scope is being drafted

If the project lead pushes back on keeping history, ask for the retention policy that supports deletion. If none exists, the default should be keep.

Common mistakes in ERP cleansing

MistakeWhy it hurtsFix
Deleting vendors with no activity in N yearsOrphans old invoices and payment historyMerge or deactivate; keep the ID
Trimming attachments to save storageLoses the scanned contracts and approvals that explain transactionsArchive attachments by location before trimming
Cleaning before the crosswalk existsNobody can trace an old ID to a new oneBuild the old-to-new ID map first
Treating the migration vendor's default scope as the client's policyDefaults favor speed, not retentionHave the controller approve scope in writing
Assuming the legacy system stays availableLicenses lapse and exports become impossibleConfirm export rights and a named exporter before cutover

When this is not worth raising

Skip the licensing angle if the company has fewer than 50 full-time employees at peak, if the ERP was implemented recently with little prior history, or if the records are mostly someone else's data held without consent. Also skip it if no one with authority over the records is reachable. In those cases, the checklist is still good cleansing practice.

Next step

Fractional CFOs and CAS leads who run migrations see systems and histories that most advisers never do. Read the fractional CFO partner overview, see how CAS growth at large firms shapes client conversations, and keep the monthly lender reporting package in view when retention holds apply. Before a handoff, the records handoff checklist and the chart of accounts redesign guide cover the same question from the other side. When a client looks like a fit, register as a partner.

The partner earns 25% of the eligible platform fees SourceX actually collects from the referred company's licensing deals, up to $100,000 cumulative per referred company, and it becomes payable only after the buyer pays and SourceX receives its fee. Rewards are not guaranteed, and the reward is never deducted from what the company receives.

  1. Step 1Share your linkSend your personal link to a company you know.
  2. Step 2Company appliesThe company applies itself at /apply.
  3. Step 3Buyer selects and paysThe buyer selects and pays for the data and SourceX receives its fee.
  4. Step 4You get your rewardYour share of SourceX fees becomes payable.

Common questions

Is it safe to delete inactive vendors during ERP cleansing?

Deactivate or merge them rather than delete. Closed invoices, payments and 1099 history point to the vendor ID, so removing the vendor can orphan those records. Keep a crosswalk of old and new IDs, and check any retention hold from your auditor, lender or tax adviser before deleting anything.

How many years of ERP history should we keep before migration?

There is no single answer; it depends on retention policy, audits, lenders and tax positions, which your own advisers must confirm. As a working rule, do not shorten history just to speed the migration. Longer, connected histories also help if the owner later explores a data license.

Does cleansing master data change who owns the records?

No. The company created the records and keeps ownership. Cleansing only changes structure and quality. Whether the data can be licensed depends on contracts, privacy promises and the company's rights, which are reviewed by the company and SourceX, not by the referring partner.

What should a CFO capture about the ERP without opening records?

Capture metadata only: the system name, module list, earliest transaction date per module, connected systems, retention settings, who can export, and where attachments live. That list is enough for a preliminary fit screen and never requires exporting or describing a single confidential record.

Who gets paid if a cleansing client later licenses its data?

Only a partner who made a valid, first introduction that led to a verified company application can earn a reward, and it is paid only after the buyer pays and SourceX receives its fee. Licensed professionals should check their own rules on referral fees and disclosure first.

Free resources

By SourceX Partnerships Team · Published 2026-10-09 · Updated 2026-10-09

Know a US company with valuable proprietary data?

Become a referral partner from anywhere we support, get your link and introduce an owner or authorized decision-maker.

Refer a company →

I own a business

Explore licensing your company's data to AI developers worldwide. Start a short assessment; no uploads needed.

Start an assessment