CRM hygiene: how to clean the database without throwing out gold
2026-03-03 · 4 min read · Adam Axelsson, founder of Revexa
CRM hygiene is about three things: merging duplicates, verifying or removing dead email addresses and updating decayed contact data. What it is not about is deleting dormant leads. A lead that has been quiet for a year is a paid asset, not junk. Here is the process that cleans the database without throwing out the gold.
The background: according to Apollo, 22.5 percent of all contact data decays per year. People change jobs, companies change domains, roles disappear. A database no one has cleaned in two years is therefore nearly half out of date. At the same time each lead costs around 100 USD to generate. That is why the cleanup should be surgical, not a mass deletion.
What should you actually remove?
Only what is objectively worthless:
- Hard bounces. Addresses that return a permanent error. The person is no longer at that address.
- Obvious junk contacts. Test records, info@ addresses without a name, competitors who downloaded a guide.
- Duplicates. Merge, do not delete. Keep the record with the most history.
- Contacts who requested deletion. GDPR gives you no choice here.
Everything else, including leads that have not replied in 14 months, stays.
Why should you not delete dormant leads?
Two numbers settle the question. According to MarketResearchFuture, 70 to 80 percent of all leads in a CRM never get followed up. And according to Forrester, 80 percent of leads stamped "not ready" buy something within 24 months, just usually from someone else. A dormant lead is therefore, as a rule, not a dead lead, it is an unworked lead.
Deleting them means you pay around 100 USD per lead again, the next time you need pipeline. Segment instead: tag leads with last activity 4 to 18 months back as "dormant" and treat them as their own list to work, not as noise to clear.
How do you find and merge duplicates?
- Match on email address first. An exact match on the address is almost always the same person.
- Then match on name plus company domain. Catches people with multiple addresses.
- Keep the record with the most history and move activities, deals and notes over before you delete the copy.
- Fix the source. Duplicates usually arise in forms and import flows. Set a duplicate check at intake, otherwise you are back at square one in six months.
Both HubSpot and Pipedrive have built-in duplicate tools that handle steps 1 and 2.
How do you handle decayed data?
Work in layers by the age of the last activity:
| Age | Action |
|---|---|
| 0 to 4 months | No cleanup, normal follow-up |
| 4 to 18 months | Tag as dormant, verify as needed, work |
| Older than 18 months | Verify email and role before every send |
Verifying means, in practice: run the addresses through a validation tool, check the title against LinkedIn or a company register on the most valuable contacts, and update the fields instead of guessing.
How often should you clean?
Quarterly is enough for most companies with 11 to 100 employees. More important than the frequency is that someone owns the task. A database without an owner decays at the same rate no matter how ambitious the last big cleanup was.
A good quarterly routine takes half a day: run the duplicate tool, export hard bounces, update the tag for dormant leads and check that new leads get an owner.
What do you do with the dormant list afterwards?
The cleanup is the prep, the work is the point. A verified list of dormant leads is the cheapest pipeline material you have, because the acquisition cost is already paid. Work it signal driven with short personal emails instead of mass sends. If you want to run the numbers on what your own dormant list is worth, there is Revexa's calculator.
Common questions
Should you delete leads that never replied?
No, not if the address is valid and the person matches your target group. Quiet leads are unworked, not dead. Only delete hard bounces, junk and deletion requests.
How much of the database is out of date?
Count on 22.5 percent decay per year, per Apollo. After two years without cleaning, roughly 40 percent of the contact data is wrong on some point.
What does poor CRM hygiene cost?
Three things: bounces that damage the domain reputation, sales time spent on the wrong people, and rebuying leads that already exist in the database at around 100 USD each.