Why this matters
Every guide in this series ends at the same dependency. Dashboards need filled fields. Forecasts need honest close dates. Automations need data they can trust. Hygiene is not a chore next to the real work. It is the load-bearing wall under all of it.
Decay is the default state. Contacts change jobs, companies get bought, deals die without anyone marking them lost, and every import adds its own sediment. An account left alone for a year is measurably wrong in every report, and the team can feel it before anyone can prove it. The feeling is what kills adoption: reps stop maintaining what they no longer believe, which accelerates the decay. A routine breaks that spiral for the cost of one hour a month.
The design decisions that matter
Decision one: one owner. Name the person who runs the monthly pass. Reps keep their own deals honest, the owner keeps the account honest. Without a name on it, hygiene happens exactly once, during the implementation.
Decision two: lost reasons as a taxonomy. Dead deals must go to lost, and lost needs a short option list: price, competitor, no decision, timing, bad fit. Free-text lost reasons produce a graveyard nobody can learn from. This list feeds the quarterly loss review, and it is a five-minute setup.
Decision three: import rules. Most mess arrives in bulk. Require that every import goes through the data owner, matches on email and organization domain to prevent duplicates, and lands cold lists in the lead inbox, never in the pipeline.
Decision four: what clean means, in filters. Define health as a set of saved filters that should return zero or near zero: deals with close dates in the past, open deals without a next activity, people without an organization, key fields empty on open deals. What you can filter, you can fix. What you can only sense, you argue about.
A worked example
The monthly routine we install, run from saved filters in a fixed order. One hour, same day each month.
| Minutes | Pass | Action |
|---|---|---|
| 0 to 10 | Duplicates | Run merge duplicates for people and organizations, merge or dismiss each pair |
| 10 to 25 | Dead deals | Open deals with no activity in 45 days: push owners to close or revive |
| 25 to 35 | Close dates | Expected close dates in the past: every deal moved or closed |
| 35 to 45 | Field gaps | Key fields empty on open deals: assign the fills to owners |
| 45 to 55 | Orphans | People without organizations, deals without contacts: relink |
| 55 to 60 | Log | Note counts per pass, compare with last month |
The log at the end is what makes it a system. Counts trending down means the habits upstream are working. A pass that grows every month points at a broken process, not at a lazy team, and that is a finding worth escalating.
Common mistakes
The heroic one-off is first. A big cleanup weekend, everything shines, and eight months later the account is back where it started, plus cynicism. Hygiene is a routine or it is theatre.
Second, mass deletion. Deleting stale deals feels decisive and destroys the win-loss history that reporting and probability calibration feed on. Lose deals with reasons, archive leads, deactivate products. Delete only junk and data you are legally required to remove.
Third, cleaning around a merge instead of doing one. Two half-maintained accounts after an acquisition need consolidation, not parallel hygiene. That is its own project, covered in the consolidation guide.
Maintenance
This page is the maintenance chapter for the whole series, so the meta-routine is short. Monthly: the hour above. Quarterly: add the field audit and a review of lost reasons, and recheck that import rules survived contact with new hires.
Yearly: one deep pass on contacts, archiving people gone silent for two years and refreshing the ones that matter. Track one number over time, the total across your health filters. When it stays near zero for two quarters, you have what most companies claim and few have: a CRM the team believes. That belief, not the software, is the asset.
Questions
Does Pipedrive detect duplicates automatically?
For people and organizations, yes. The merge duplicates feature finds likely pairs by name, email and phone, and lets you merge them with a chosen surviving record. Deals have no automatic detection, so those you find with filters.
What happens to history when two records merge?
It combines. Deals, activities, notes and emails from both records end up on the survivor. For conflicting field values, the record you keep as primary wins, so check the fields before confirming.
How long should a monthly hygiene routine take?
One focused hour for an account under 50,000 records, once the backlog is gone. The first cleanup is the exception and can take days. That is the price of the years nobody did this.
Should reps clean their own data or should one person own it?
Both, split cleanly. Reps own their deals: close dates, stages, dead deal decisions. One data owner runs the account-wide passes for duplicates, imports and field audits. Hygiene owned by everyone is owned by no one.
Is deleting old data ever the right call?
Sometimes, and GDPR may even require it for personal data with no business purpose. But default to archiving and losing deals with reasons. Delete personal data on retention schedule, not history you may need.