Close the quarter by counting the exceptions

Most quarters end with a demo and a feeling. The demo shows what shipped; the feeling — pretty good, rough one, we'll catch up — becomes the organisation's memory of the quarter. Neither is data. A quarter actually closes on three counts, and each one changes something specific about the next quarter's plan.

Count one: honest closes

Of the commitments made, how many reached a recorded outcome — delivered clean, delivered with a named difference, or not delivered — each closed with a verdict and a written lesson. A commitment that just stopped is not closed; it is abandoned, and it counts against this number, not toward it. This is the count a completed-and-learned view exists to answer at a glance, and it is the numerator of your organisation's trust: the ratio of promises to honest closes is the only delivery metric leadership can verify for itself.

What it changes: the size of next quarter's line. A group that closed eight of thirteen honestly does not need better estimates; it needs to commit less until the ratio recovers. Committing eleven and closing eleven builds more credibility in two quarters than any dashboard ever has.

Count two: slips, with the differences named

Of the non-clean closes, what actually differed from the promise — not the fact of slipping, which count one already holds, but the substance: what shipped instead, what was cut, what moved and why. This count only exists if closes record the difference, which is why a close should never be a checkbox.

The value is in the clustering, because differences are rarely random. Three slips that all trace to an unconfirmed dependency point at the planning stage. Differences that trace to capacity optimism point at the estimates and their confidence levels. Scope that grew mid-quarter points at the door. The cluster names the stage that is lying to you — which is the one thing a "we missed some dates" retrospective can never do.

Count three: early-commitment exceptions

How many of the quarter's commitments crossed the readiness bar on a recorded exception — committed early, with why-now, what's-missing, who-accepts and the changing condition written down at the time. This count is the reason the exception exists at all: an early-commitments report is only possible because every early commitment paid its four fields on the way in.

It is also the most diagnostic of the three. Line count three up against the misses. If most of your misses were early commitments, your estimation is fine and your door is broken — dates are being promised before evidence, and the fix is upstream of every retro action item you were about to write. And if the exception count itself rises quarter on quarter, escalation has become cheaper than the front door, and the price needs rebalancing.

A worked example

Nadia runs delivery for a sixty-engineer product group. Her Q3 close: thirteen commitments. Eight closed clean. Three closed with named differences — one scope exclusion dropped late, one three-week slip, one delivered without its migration. Two not delivered. All thirteen carry a verdict and a lesson, so count one reads: thirteen of thirteen closed honestly, eight clean. The differences cluster hard: both non-deliveries and the slip trace to one platform team whose "confirmed" dates were never actually confirmed. Count three: four commitments were early exceptions — and three of the five non-clean closes were among those four.

The counts write next quarter's changes almost by themselves. The line drops from thirteen to eleven. Nothing that depends on the platform team gets committed without a date that team recorded itself. And the escalation path that produced three of the four early commitments now runs through the planning room, out loud, at the same four-field price. None of this came from the demo, and none of it would have survived as "the feeling".

The counts are the report

Conveniently, the three counts are also the leadership report — they are exactly the shape of promises against outcomes, compressed to one screen. A leadership team that hears the same three counts every quarter learns to trust the line behind them, and a leadership team that trusts the line stops asking for the activity lists. What happens between the commitment and the close — status that carries reasons, outcomes that are comparisons — is the subject of the delivery overview.