CRM migration mistakes: five ways it goes wrong and how to avoid them

What goes wrong in a CRM migration?

You’ve picked a new CRM and you’re ready to move your contacts, deals, and history across. This is the moment most projects quietly go wrong, and almost never because of the software.

A CRM migration is the process of moving your customer data from an old system (a spreadsheet, an inbox, or a previous CRM like HubSpot or Salesforce) into a new one, along with the structure and history that give that data meaning. CRM migrations fail for three reasons that have nothing to do with which platform you chose: the data going in is messy, the fields and history don’t map cleanly across, and the team is never properly brought along. Get those three right and switching CRM is straightforward. Get them wrong and you carry every old problem into an expensive new system.

This guide covers the five mistakes that cause the most damage and gives you a pre-migration checklist to avoid them. It’s the risk-and-failure companion to our step-by-step migration guide, which walks through the full process in order.

Contents

• Do you actually need to migrate all your data?
• Mistake 1: Importing dirty data
• Mistake 2: Broken field and history mapping
• Mistake 3: Migrating everything instead of what matters
• Mistake 4: Building reports on un-deduplicated data
• Mistake 5: Forgetting the team
• Your pre-migration checklist
• TL;DR

Do you actually need to migrate all your data?

Before planning any migration, ask an honest question most guides skip: do you need to bring all of it across?

A surprising amount of what lives in an old CRM or spreadsheet is dead weight. Contacts you have not spoken to in five years, deals that closed in 2019, duplicate companies, and notes attached to the wrong person. Migrating that data doesn’t preserve value, it preserves clutter, and it slows the import down.

A good rule of thumb: bring every active contact and company, the last 12 months of won and lost deals, and any historic notes that genuinely inform a current relationship. Archive the rest to a backup file rather than importing it. You keep the record if you ever need it, without polluting your shiny new system on day one.

Mistake 1: Importing dirty data

This is the single most common and most damaging mistake we see businesses make. If you export a messy contact list and import it straight into a new CRM, you haven’t fixed anything. You’ve just moved your mess to a more expensive location, faster!

Dirty data means duplicate records, multiple names crammed into a single field, inconsistent company naming (Acme Ltd, ACME Limited, Acme Co.), missing email addresses, and dates in three different formats. In our experience cleaning SMB databases before migration, somewhere between 10 and 25 percent of records typically carry a duplicate, a formatting error, or a gap that needs attention before import. Most CRMs reject organisation records outright when they detect duplicate company names, so a dirty file doesn’t just look bad, it breaks the import itself.

The fix is to clean the data at source, in the spreadsheet, before you touch the CRM. Deduplicate, split full names into separate columns, standardise company names to one agreed version, and validate that every email contains an @ and a domain. Assign one person to own this, as cleaning data by committee rarely ends well.

Mistake 2: Broken field and history mapping

Your old system and your new one do not store data the same way. A field called “Lead Source” in one CRM might be a tag in another, and the two aren’t interchangeable, so it pays to be deliberate about planning your tags and fields before you import. Custom fields, deal stages, and note histories are where exports most often break, and where data silently disappears if nobody is checking.

This is especially true when migrating from HubSpot or Salesforce, where years of activity history, custom properties, and pipeline stages have built up. A raw CSV export rarely carries all of that across intact. Custom fields land in the wrong place, dropdown values stop matching the new picklist, and note timestamps get lost, leaving a contact record that looks present but has lost its memory.

The table below shows what typically exports cleanly from each common source, and what tends to need extra attention.

The fix is to map your fields explicitly before importing, then run a test migration with a small sample (50 to 100 records is plenty) and check the result by hand. It’s far easier to spot a mapping error on 100 records than to untangle it across 5,000 after the fact.

Mistake 3: Migrating everything instead of what matters

Closely related to importing dirty data is importing too much of it. Teams often treat a migration as a chance to preserve every record “just in case”, and end up importing a decade of dead contacts alongside the live ones.

The problem isn’t just clutter. A new CRM full of stale, unworkable records is harder for your team to trust and adopt. When half the contacts are people nobody recognises, the system feels unreliable, and people quietly drift back to their own spreadsheets.

Decide your cut-off before you export. Active contacts and companies come across. Recent deals come across. Genuinely useful history comes across. Everything else goes to an archive file you keep safe but do not import. This is the same discipline that stops a new CRM feeling like the old mess in new clothes.

Mistake 4: Building reports on un-deduplicated data

This mistake shows up weeks after the migration, when someone runs a report and the numbers look wrong. If duplicate contacts and companies made it through the import, every count, every conversion rate, and every pipeline total is quietly inflated or distorted.

It’s worse than no data, because it looks authoritative. A pipeline report built on a database where 15 percent of companies appear twice will overstate your opportunity count and understate your win rate, and you’ll make decisions on figures that were never real. The same problem undermines any AI feature you layer on top later: a summary or a forecast built on messy data is just a faster way to be wrong.

The fix happens before the import, not after. Deduplicate at source, and after the full migration validate a representative sample (around 5 percent across different contact types) against your source counts before you let anyone build a single report. Validation is the step that confirms the migration actually worked, rather than assuming it did.

Mistake 5: Forgetting the team

A clean, well-mapped migration still fails if the people who use the CRM every day aren’t brought along. This is the quietest failure mode and the most common reason a technically successful migration ends in a return to spreadsheets within six months.

Preparing your team for a CRM migration isn’t a one-line email with a login link. It means showing the people who will use it most how to do the specific things they do every day in the new system: logging a call, moving a deal, finding a contact, adding a note. It means written guidance specific to your setup, not generic vendor help articles. And it means a review session two to four weeks after launch, by which point real situations have surfaced that the first walkthrough never covered.

Treat the two-to-four week review as a ceiling rather than a target. The sooner you catch the workflow questions and the small frustrations, the less chance they harden into workarounds that route around the CRM entirely.

We saw the payoff of getting all of this right when we migrated this agency off HubSpot: years of relationship history were preserved intact, costs dropped by around 60 percent, and lead capture was automated rather than re-keyed by hand. The migration worked because the data was cleaned and validated first, not because the destination CRM was magic.

Your pre-migration checklist

Run through these before you export a single record:

If your data sits in several places, your previous CRM doesn’t export cleanly, or your first attempt has already gone wrong, this is the point where professional help is usually faster and cheaper than another DIY round. The same is true if you are leaving an over-priced platform: our guide to switching CRM from HubSpot covers the timing and the trade-offs, and the true cost of staying on spreadsheets makes the case for moving at all.

TL;DR

• CRM migrations fail on data, mapping, and people, not on the software.
• Clean the data at source first. Importing dirty data just moves your mess somewhere more expensive.
• Map custom fields, deal stages, and history explicitly. HubSpot and Salesforce exports rarely carry everything across intact.
• Migrate what your business will use, not everything. Archive the rest rather than importing it.
• Deduplicate before import, then validate a sample afterwards. Reports built on duplicate records are worse than no reports.
• Bring the team along with a real walkthrough and a review session two to four weeks in.
• Run the pre-migration checklist before you export a single record.