CRM Consolidation After an Acquisition: Model, Duplicates — Odoo & ERP article by Rui Codex

After an acquisition, consolidate several CRMs by first agreeing one shared data model and ownership rules, then migrating records in controlled stages, and only then merging duplicates. Doing these steps in a different order usually causes lost history, conflicting fields and unclear responsibility for customer data.

Key takeaways

  • Agree on one data model and field definitions before moving any records.
  • Match duplicates on rules you can explain, and keep a record of every merge decision.
  • Assign a named owner to each data domain, not only to the system.
  • Treat GDPR obligations as part of the migration plan from the start.

Why do acquired companies end up with several CRMs?

Each company keeps the customer system it already used, so after a deal the combined group often runs two or more CRMs side by side. Sales teams keep working in their own tool, marketing lists diverge and finance cannot see one view of a customer.

Running parallel systems is sometimes a sensible short-term choice while contracts end and integrations are built. It becomes a problem when no one decides the end state, because duplicate and conflicting records then accumulate every week.

What should the target data model look like?

The target data model is the set of entities, fields and relationships that the combined business will use, and it should be agreed before any migration tooling is written. Start with the core objects: accounts (companies), contacts (people), opportunities, activities and products. For each, write down the meaning of every field, its allowed values and who is allowed to change it.

Most conflicts come from fields that look identical but mean different things. One system may store a customer status as a free-text label, while the other uses a lifecycle stage with fixed steps. Map these differences explicitly and decide the canonical version rather than copying both.

Map fields by meaning, not by name

  • List every field in each source system with an example value.
  • Mark fields that carry the same meaning, fields that differ, and fields that exist in only one system.
  • Choose a canonical definition and keep the original value in a legacy field for audit purposes.
  • Record the relationship rules, such as one account having many contacts, and check them against real data.
Tip: Export a sample of two hundred real records from each system before designing the model. Edge cases, such as contacts without a company or accounts with several legal entities, appear quickly in real data and rarely in the design documents.

Where the systems use different currencies, time zones or identifiers, decide the standard now. Changing the identifier scheme after records are merged is far harder than choosing it at the start. A reasonable starting point is one stable internal ID per record that never changes, with the old system IDs stored as references.

How do you find and merge duplicate records?

You find duplicates by comparing records on rules you can explain, and you merge them only after a person has reviewed the rules and the exceptions. Exact matches on email address or tax number are safe to automate. Fuzzy matches on company names or phone numbers need review because two different companies can share a name.

Merging is destructive in practice, so keep the source records and a log of each decision. A merge that cannot be reversed will eventually produce a dispute about which activity history belongs to which account.

Trade-offs of automatic and manual merging

Automatic merging is fast and suits large volumes of exact matches. It fails when the match rule is too loose, because it silently joins unrelated customers, and that error spreads into invoices and reports.

Manual review is slow but accurate for ambiguous cases. It fails when reviewers have no context, so give them the source system, last activity date and owner for each candidate pair.

Who should own records after the merge?

Each data domain should have one accountable owner, usually a business role such as the sales operations lead for account data and the finance lead for billing contacts. The owner decides field definitions, approves merge rules and resolves disputes. Without this role, the CRM becomes a place where everyone edits and nobody is responsible.

Ownership also covers the records themselves. Decide which team owns an account after the merge when two sales teams have worked the same customer, and write that rule down before the migration, not during it.

Tip: Set up a short monthly data review with one representative from each former company. Use it to approve rule changes and to clear the queue of unresolved duplicates.

What is a safe migration sequence?

A safe sequence moves one data domain at a time, validates it, and only then retires the source system. Begin with reference data such as products and owners, then accounts, then contacts, and finally activities and open opportunities. Each step should have a rollback plan.

  1. Freeze the data model and the ownership rules, and publish them to all users.
  2. Clean the source data: remove obvious junk, standardise formats and fix broken references.
  3. Load a test migration into a sandbox and compare counts and sample records with the sources.
  4. Run the duplicate process with logging, then review exceptions with the domain owners.
  5. Switch users in groups, keep the old system read-only for a defined period, and then archive it.

Integrations are the usual failure point. If a marketing tool or ERP still writes to the old CRM after the switch, two versions of the customer reappear within weeks. Inventory every connection before cutover, and treat the connection list as part of the project scope. Teams that need the sales and finance systems to share customer records can review how ERP and CRM integration is handled in a typical project before they set the order of work.

Hypothetical examples

Here are two hypothetical scenarios:

Scenario one. A services firm acquires a smaller competitor. Both use different CRMs, and the same client appears under two company names. The team agrees that tax number is the match key, reviews the name-only matches manually, and keeps the acquired firm's activity history attached to the surviving account.

Scenario two. A distributor acquires a manufacturer whose CRM stores product interest as free text. The team maps those texts to a fixed list of product families, keeps the original text in a legacy field, and assigns the product owner to approve any new family. Sales reps see one list after the switch, and nobody has to re-enter history.

Neither scenario describes a real organisation. They show the order of decisions that tends to work: model first, match rules second, ownership third.

How does GDPR affect merging customer data?

Under the GDPR, personal data must be processed lawfully, and Article 5 sets out principles including data minimisation and accuracy. Merging contacts from two companies is processing, so the combined business needs a documented purpose for keeping each record and must not retain data it no longer needs. Read the regulation on EUR-Lex, the official EU law database.

This article is general information and not legal or tax advice. Rui Codex offers technical implementation support for CRM migrations, but it does not provide legal, tax or compliance advice, audits or certification. Involve your data protection officer or qualified counsel when you decide retention periods and lawful bases.

In practice, the migration should carry the consent and purpose data with each contact, and deletion requests received during the transition must be applied to every copy, including archives. If you need a technical review of how the systems will handle those requests, you can request a project consultation, and the system integration service page describes the technical scope.

Frequently Asked Questions

Should we migrate everything from the acquired CRM?

No, you should migrate only the records and history that the combined business needs for current operations and legal retention. Old closed opportunities and unused contacts can often stay in an archive that is read-only. Document the selection rule so that it can be explained later to auditors or to the people whose data was moved or left behind.

How long should the old CRM stay running?

Keep the old CRM in read-only mode for a defined period that covers open contracts and reporting cycles, then archive it. The exact length depends on your contracts and internal reporting needs. Set a retirement date at the start of the project so that the parallel period does not extend indefinitely.

Who decides which duplicate record survives?

The domain owner decides which record survives, using the agreed survivorship rules. Typically the record with the most complete and recent data is kept, and the other record's history is attached to it. Each decision should be logged with the reason so that disputes can be resolved later.

Can a CRM migration be done without a data model?

A migration without a data model is possible, but it usually moves inconsistencies into the new system. Fields will be copied as they are, and conflicting meanings will persist. Agreeing the model first costs a few workshops and prevents rework during cutover and reporting.

Where can we see examples of delivered integration work?

Published case studies are available on the Rui Codex case studies page, which describes the projects that have been documented. Read them for the scope and approach. For a discussion of your own systems, contact the team by email at info@ruicodex.com.

Sources

Tags: CRM consolidation data model duplicate records data ownership CRM migration ERP and CRM integration GDPR M&A integration

Need Help Implementing This?

Our team can help you put these insights into practice. From AI automation to custom software development, we build solutions that deliver real results.

Send a project request