Neo D.

In the small teams I've seen, it's almost always the periodic sweep — monthly, sometimes quarterly — and it's usually whoever owns ops doing it manually. Write-time rules catch the obvious stuff, but they only cover the paths someone thought of, which is exactly the point Dubray made earlier. Nobody at five reps is running real-time drift detection. That tooling exists, but it's the expensive enterprise category and it's overkill at that size. So the sweep is the safety net, and it's the part that quietly costs the most. It's where you find the same account under three spellings, a batch of records missing the field your reports depend on, phone numbers in four formats. In practice it usually starts with someone exporting the CRM, cleaning it in a spreadsheet, and reimporting it — which is the part I work on. The teams that made it stick were the ones where the sweep was quick and boring. When it takes a day, it gets skipped. When it takes an hour, it happens.

One thing I'd add from the data side: reps stop updating when they stop trusting what's in the CRM. If the pipeline shows the same account under three spellings, or half the records are missing the field the reports depend on, the CRM stops being the source of truth — it becomes a chore. The teams I've seen hold the habit fixed the data quality first, so the CRM actually reflected reality, and then the daily update felt like maintaining something useful rather than feeding a broken system. The screenshare point is real, but I think it only works if what's on screen is trustworthy. Showing a pipeline full of dupes and blanks doesn't shame people into updating — it just confirms the thing isn't worth updating.