How RevOps teams actually define a stale deal

Finding 1

Nobody agrees on what "stale" means

Days in stage was the most common answer, but it was rarely the one people defended when questioned. Several practitioners argued that time in stage is close to meaningless on its own — an enterprise deal sitting 30 days in legal review is healthy, and a three-day-old deal with no next step booked is already in trouble.

The signals people actually named, roughly in order of how often they came up: no activity logged, no next step scheduled, a close date that keeps getting pushed, the champion going quiet, and the deal simply being older than that stage's norm.

Finding 2

Close-date slippage is the signal almost nobody tracks

A close date that moves once is normal. A close date that's moved four times is a deal nobody wants to call dead. Several people described recognising this pattern manually — and none described a system that caught it automatically.

The reason is structural: a CRM stores the current close date, not the history of what it used to be. The number looks perfectly valid every time you check it. The story is in the movement, and the movement isn't in the field.

Finding 3

Detection wasn't the hard part. Enforcement was.

Most teams could produce a list of stale deals if you asked them to. What kept coming up instead was what happens next — or rather, what doesn't.

One practitioner described a genuinely well-designed escalation ladder: notify the rep at one week, the manager at two, leadership at three, close it at four. When asked whether reps actually behaved differently at week two versus week three, the answer was less certain. The ladder existed. Whether it changed anything was an open question.

This gap came up more than any single detection method: the alert fires, everyone sees it, and the deal stays exactly where it was.

Finding 4

Every good system had an escape hatch

The teams whose processes had actually survived more than a quarter all had one thing in common: a way to legitimately exempt a deal.

One team used an account-level tag as a manual override, specifically so the automation wouldn't keep nagging about a deal that was genuinely paused for a good reason. It was described as something that only became obvious after being burned by a rule too rigid to accommodate reality.

Rules without exceptions don't get followed. They get ignored, and then the whole system quietly stops meaning anything.

Finding 5

A green dashboard is not the same as a healthy pipeline

The most uncomfortable version of this problem isn't a deal that looks bad. It's a pipeline that looks fine.

One RevOps lead described discovering that a broken integration had silently marked follow-up tasks complete across their entire pipeline. Nothing errored. No alert fired. The board was green for two weeks, and the problem only surfaced when a customer emailed the CEO asking whether they'd been forgotten.

Most CRM setups are built to trust their own status fields completely. Very few have anything independently checking whether the appearance of progress still corresponds to something actually happening.