How the diagnosis works

Seven things that shape what we check for and why each one matters. If you want to run any of this yourself, the descriptions here are enough to do it.

Most companies compute average sales cycle from deals that closed.

That excludes every deal still open. Deals stay open because they stalled, so the ones you drop are disproportionately the slow ones. The number that comes back is faster than reality, and it gets faster every time something gets stuck.

The fix is to compute against every opportunity created in the window, carrying open deals at time elapsed rather than dropping them. In SQL that is the difference between an inner join and a left join. One line.

Both numbers are correct. They answer different questions. The one on your slide answers how long a deal takes if it closes. The one you need answers how long a deal takes.

This matters more here than in most industries because the cycles are long. A six month gap between the two numbers changes your hiring plan and your cash forecast.

Every company is running three revenue engines at once.

The distance between the three is where most findings come from.

The evidence for the third one is not hidden. It is personal spreadsheets, self-built dashboards, Slack threads used as a system of record, and manual reconciliation before every report.

People route around a system for a reason. Usually the official version does not fit the work. So the workaround is information, not a violation.

Every company fails at one of three layers. Which one tells you what to fix.

Failure demand is work that exists only because something failed the first time.

Re-entering data that arrived wrong. Rebuilding a report because the numbers looked off. Rerouting a record that went to the wrong person. Rescuing a deal that went quiet because nobody was watching. Calming a customer who should not have been surprised.

None of it appears on anyone's job description. All of it is on someone's calendar.

This is how we count what a leak costs. We count hours spent on work that should not have existed. The work is observable and the cause is nameable.

A busy team is not necessarily a productive one. Sometimes it is a team processing the consequences of a broken system.

An average across two populations that behave differently describes neither of them.

Your win rate is a blend of a utility deal and a corporate deal. Your cycle length is a blend of a repeat customer and a first-time buyer. When the blend moves, nobody can say whether performance changed or the mix did.

The test has four steps.

This works on win rate, cycle length, CAC, and retention. Split by segment, by source, by deal size, and by rep tenure.

It also settles definitional arguments. When sales and marketing disagree about what qualified means, both are correct under their own definition, which is why the meeting never resolves. Run both definitions against closed won data instead. Whichever population converts differently is the one carrying information.

Separation gives you two populations that behave differently. It does not tell you why, and the why decides whether the gap is worth anything.

One team ran the numbers and found their California accounts closed at a much higher rate. The finding was real. The explanation was not. Their reps sat on the east coast, so west coast accounts fell inside a wider calling window and got more attempts before anyone gave up. Location was standing in for how many times someone picked up the phone.

Nothing in the data says which of those it is. The gap looks identical either way.

So every separation that survives gets one more question asked of it, which is what else moves alongside it. Territory. Rep tenure. Deal size. When that segment started being sold to at all. If a third thing explains both sides, the segment was never the finding.

Any change to how deals are qualified, routed, worked, or handed off is invisible in win rate and cycle length until a full cycle has passed.

With cycles running six to eighteen months, that means an intervention made in January shows up in win rate somewhere in the second half of the year. Anything you measure before then is noise.

Five things do move immediately. Opportunity creation rate, qualification pass rate, first touch percentage, time to first touch, stage entry counts.

We state this at the start of an engagement, not at the end. Otherwise good work gets judged on a number that could not have moved yet, and the intervention gets reversed for the wrong reason.

One thing that follows from it. Tightening qualification lowers pipeline volume. That is the mechanism working. Volume down and revenue up is the success case, and it reads as failure to anyone who was not told in advance.