RevOps Strategy
August 27, 2026 · 5 min read · by Ananda Narasimhan
Ask a VP of Sales how confident they are in their CRM data and most say "pretty confident." Ask them to pull the actual open pipeline number from three different reports and watch the confidence disappear. The CRM was never lying. It was just recording whatever got typed into it, by whoever remembered to type, whenever they got around to it.
A hygiene audit is not a cleanup project. It is a diagnostic that tells you whether the numbers your team reports on are real, and it takes about a day to run properly.
Pull a duplicate report on contacts, companies, and deals using both exact-match and fuzzy-match logic. Exact match catches the obvious case: same email, two records. Fuzzy match catches the expensive case: "Acme Inc." and "Acme, Inc" as two separate companies, each with its own deal, each showing up separately in a pipeline report that leadership assumes is deduplicated. A 10-15% duplicate rate on companies is normal for a CRM that has never been audited. Above 20%, your pipeline and ARR numbers are wrong by more than rounding error.
Pull 50 records currently marked "Customer" and check whether they actually have an active, paid subscription. Pull 50 marked "SQL" and check whether a rep actually qualified them, or whether the stage moved because a workflow rule fired on a form fill. Lifecycle stage drifts the same way lifecycle stage definitions drift between systems: quietly, one exception at a time, until the field describes what a workflow did six months ago instead of what is true today.
Every open deal needs an owner who is still employed, still covering that territory, and still able to act on it. Ownership rot is one of the most common findings in an audit: reps who left six months ago still own 40 open deals, and nobody reassigned them because reassignment isn't anyone's job until a QBR forces the question. Pull every open deal, group by owner, and cross-reference against your current headcount. Anything orphaned needs a new owner today, not at the next reorg.
Run a fill-rate report across every custom field on your key objects. Fields below 20% fill rate are either not required, not understood, or not useful, and each one is a maintenance cost with no return. On the fields that are filled in, spot-check for garbage values: a lead source dropdown with 40 options because reps kept adding "Other: ___" instead of picking an existing one, a company size field where half the entries are obviously copy-pasted from the wrong row. Trim the unused fields. Fix the dropdown before adding a 41st option.
Pull every open deal with a close date in the past. This number should be close to zero and almost never is. A stale close date means a rep stopped updating the deal weeks or months ago, which means the "Best Case" and "Commit" numbers built on top of it are fiction. This single check is usually the fastest way to show a sales leader that the forecast they've been presenting has more drift in it than they assumed.
A CRM with genuinely trustworthy data has a duplicate rate under 5% on core objects, lifecycle stage definitions that match a written definition your team could recite, zero orphaned deal owners, a handful of well-used custom fields instead of forty half-used ones, and no open deals with a close date in the past. Most CRMs we audit hit one or two of these. None hit all five without a recurring process behind them.
That process does not need to be complicated. A monthly duplicate scan, a quarterly field audit, and an ownership check every time someone leaves the sales team catches nearly everything before it compounds. The CRM does not need heroics. It needs someone treating data quality as an owned, recurring responsibility instead of a project you run once and hope holds.