Deduplicating first feels like progress. The duplicate count drops every day, and someone can report a smaller number at the Monday meeting. Fixing stage, lifecycle and owner definitions first is slower, and it is the only order in which the merge decisions come out right.
If you have room for one HubSpot CRM cleanup before January, this is the order we would use, and the reason duplicates go last.
Why duplicates should not be first
Every merge is a decision about which values survive. According to HubSpot’s guide to merging records, the primary record’s values win, with a few exceptions. One of them matters here: on merged contacts and companies, the lifecycle stage becomes whichever of the two records was further along the funnel.
The second fact is harder. HubSpot’s own documentation says merged records cannot be unmerged. So if “Customer” means three different things in your portal, every merge that touches lifecycle stage freezes one of those meanings into the record for good. If nobody agreed which owner keeps an account, the merge makes that choice for you, one pair at a time.
There is also the question of where duplicates come from. HubSpot already deduplicates contacts by email and companies by company domain name when records arrive through imports and forms. The duplicates that remain are often created by people: a second contact so a rep can open a second deal, a personal address next to a work one. Merge those without changing the rule and they come back next quarter.
The HubSpot CRM cleanup order we use
- Definitions. Write one sentence for each lifecycle stage: what it means, which event moves a record into it, and who owns the record at that point. Do the same for deal stages, with one exit criterion per stage. This is a document before it is a setting.
- Required fields. Pick the few fields that reports and routing depend on, and make them required at the moment the information is actually known. This stops new bad records while you clean the old ones.
- Duplicates. Now merge, with written rules for which record is primary and which owner wins. Our post on duplicate merge rules covers that order in detail.
Step two comes before step three for a simple reason. If nothing stops the inflow, you clean the same pile twice.
Who this matters for
Heads of sales feel this in the forecast: duplicate contacts with separate deals and separate owners make pipeline numbers look bigger than they are. Heads of marketing feel it in lifecycle reporting, where one person counted twice can be a lead and a customer at the same time.
The tools involved are not a premium extra. HubSpot lists record merging and lifecycle stage settings for all products and plans, so this cleanup is about process and time, not about buying a higher tier.
What to check in your portal
- Go to Settings, then Objects, then Contacts, and open the Lifecycle Stage tab (HubSpot’s lifecycle stage guide). Compare each stage with your written definition. Editing stages needs Super Admin permissions.
- Open Data Management, then Data Quality, and the Manage Duplicates tab. Sort a sample of pairs into accidents and records created on purpose. The second group is a process problem, and the owners can tell you which rule is missing.
- In Automation, then Workflows, list every workflow that sets lifecycle stage, owner or lead status. If your new definitions change the logic, these workflows must change too, or they will undo the cleanup overnight.
- Know your safety net before any bulk change by workflow. The workflow data restore feature can undo property changes made by a specific workflow, but only for the last 14 days, only for Super Admins, and not for custom-coded actions. Check on that page whether your edition includes it.
Our take
A high duplicate count is a symptom. In most portals we audit, it mirrors a process that was never written down. Nobody decided when a lead becomes a customer or who owns a returning buyer. HubSpot records the answer. It cannot invent one.
The definitions step is not admin work. It takes a working session between the heads of sales and marketing, and one named person who owns the document afterwards. The merging that follows is admin work, and it goes faster once every pair has a rule.
When to skip our order: if your duplicates come from one bad import, merge them and move on, because definitions will not change those decisions. And if nobody can own the definitions before January, do not start mass merging. Write the definitions first, even if the merging waits a month.
Free checklist
Run the 25-check HubSpot audit yourself
Every check comes with the exact filter and the number that means you failed.
FAQ
How do I clean up my HubSpot CRM?
Start with definitions, not data. Define each lifecycle stage, deal stage and owner rule, then make key fields required, then merge duplicates by written rules. That order stops the mess from coming back.
Should I merge duplicates before or after fixing lifecycle stages?
After. Merged contacts and companies take the lifecycle stage of whichever record was further along, so merging before the stages mean something locks in the wrong values. Fix the definitions, update the workflows that set them, then merge.
Can you undo a merge in HubSpot?
No. HubSpot’s documentation states that merged records cannot be unmerged. The workflow restore feature is about property changes made by workflows, so do not count on it for merges. Agree on the merge rules before anyone starts clicking.
This week, book one working session with your sales and marketing leads and write the stage and owner definitions on a single page. Then open the Manage Duplicates tab and see how many pairs your new rules already answer. If you want a structured way to find where to start, the HubSpot audit checklist walks through it.
Related reading
Alignify
On this page
Get the audit checklist
All 25 checks as a spreadsheet with auto-scoring.
Talk to Alignify
45 minutes, your portal on screen, no slides. We show what we would fix first.