Q4 Is Built in September

The calendar does not negotiate

Arithmetic first. September 23 to December 31, inclusive, is 100 days. Take out the weekends and it is 72 weekdays. Take out Columbus Day, Veterans Day, and the two days of Thanksgiving, then the stretch from December 21 when contracts sit in inboxes behind out-of-office replies, and you are somewhere around 60 working days. Nobody argues with the calendar. Everybody argues about what to do with it.

Our position on the 60 days: none of them are for reading dashboards. A dashboard read in October describes an account that was in market in September. By the time the report renders, the season has started without you.

Q4 is built in September

We are not going to tell you your cycle length. Cycles vary by price, by segment, and by how many people have to say yes, so any benchmark we quoted would be wrong for half of you. Do the subtraction yourself. Take the last day anyone at your customers signs anything, which for a lot of teams sits around December 18, and count backward by your cycle. Sixty days lands in mid-October. Ninety lands in September. A hundred and twenty lands in August, which means those deals are already late.

The shape is the point, not the number. Whatever your cycle, the accounts that will close before the year ends are not waiting to be discovered in November. They are in market now, and being in market leaves marks: on your pricing page, inside your product, in your CRM. The September question is not "what is the Q4 plan." It is "which accounts are already acting, and who owns each one."

Figure 1. Three accounts fire in September and close in November and December. The cycle band is a shape, not a length (illustrative, not a benchmark).

Three Q4 specimens

Three patterns read differently in a spending season. We name them by their parts, the way the Field Guide does.

The renewal window. renewal T-90 → pricing revisit. A current customer whose contract ends before December 31, reading your pricing page about ninety days out. In March that is a renewal conversation. In Q4 it is the year's last renewal cohort choosing between expansion and exit, and both land in this fiscal year. Company resolution is enough; the CRM already has the names and the date. Play: the account owner gets a note with the renewal date, the current plan, and the pages read, and nothing else happens until they answer.

The budget check. pricing → plan-compare → repeat (7d) on a fit account. This is the pattern from The Class You Own, and in October it means something it does not mean in April: somebody is checking whether your number fits a budget that gets spent this year or locked for next. Company resolution is enough to route it, the rung The Class You Rent recommends; reaching a named person is a decision the Play makes, not the signal. Play: the account owner gets the pages and the score by axis, and the account goes on the Q4 watch list.

The boomerang. closed-lost → pricing return (6mo+). An account that told you no in the spring, back on your pricing page in the fall. Brad's part 3 of the Field Guide yesterday covers the whole class this comes from, so we will keep to the Q4 reading: this account already knows your name, sat through your demo, and either has budget again or has lost patience with whoever won. Company resolution is enough; the closed-lost record has the contact. Play: the account owner gets the original loss reason and the pages read, and nothing else happens until they answer.

Three patterns that read differently in a spending season. Each shows its parts, the Q4 reading, the resolution it needs, the one pre-approved action, and the record line Bryn would write.

Illustrative, not a benchmark. Accounts anonymized; scores, dates, and Play names made up to show the shape. Company-level resolution is enough for two of the three; the budget check needs company to route and a separate decision to reach a person.

Figure 2. Three Q4 specimens by their parts, with the resolution each needs and the one pre-approved action (illustrative, not a benchmark).

If you already run an agent, it can ask Bryn which of these fired this week; Monday's piece covers what your agents can ask Bryn over MCP, and "which fit accounts revisited pricing inside T-90 of renewal" is a fair question.

The working-days math

Now the hard part, which is not the signals. It is the hands.

Say the three specimens fire on a modest number of accounts between now and Thanksgiving. Each firing needs someone to see it, check it against the CRM, decide whether it is real, write the note, and remember to follow up when nobody answers. That is the labor Brad described in Follow-Up Has No Owner: not strategy, but checking, cross-referencing, guessing, writing, forgetting. In Q4 it runs across the same six tools Chris walked through on Tuesday in Your Pipeline Runs Across Six Tools, which is why the controls have to cross the tools too. A Q4 push that lives inside one tool's agent misses the signals in the other five.

Put your own numbers in. The calendar is real; the hours are yours.

The dates are the 2026 calendar, computed from real dates. The hours are yours. Everything derived from the hours is illustrative.

Working days are computed from the 2026 calendar; only holiday weekdays inside the chosen span are subtracted. Hours of follow-up, the working days they add up to, and what a Play could give back are illustrative, not a benchmark.

Figure 3. Working days left in the year against the hours a week you spend hand-running follow-up. The dates are the 2026 calendar; everything derived from your hours is illustrative, not a benchmark.

Sixty working days, with a slice of each spent hand-running follow-up that a pattern already described, is a season where the operator is the bottleneck and the signals are not.

Who owns a September signal

An owner is one of two things: a person or a Play. That was Brad's rule on Labor Day, and it holds harder in Q4 because the deadline is real.

The pipeline review is not an owner. A weekly meeting where the team looks at a list is a place where signals go to be discussed. It does not act inside the hour; it acts inside the week, if it acts at all. In a quarter of twelve working weeks, a week is a twelfth of the runway.

Do this today. Write the three specimens down by their parts. Next to each, write a name. If a person's name fits, that person owns it: they see it fire, they run the pre-approved action, they log what they did. If nobody's name fits, and for at least one of the three nobody's will, it is a Play. That is the whole decision. Brad made the case in One Operator, a Portfolio of Plays: the unit of leverage is the Play, and one operator can carry a portfolio of them through a season.

How Bryn runs the season

Bryn is not another dashboard to watch. It is the governed execution layer that runs Plays through your stack.

Bryn watches the three Q4 specimens the same way it watches any pattern. It names the pattern by its parts, scores the account against your profile (a December renewal date moves the timing axis; the pattern does not change), runs the Play you approved inside the boundary you set, and writes the record. The lines below are illustrative and the accounts anonymized; the shape is what the audit log shows.

I matched renewal T-90 → pricing revisit on acct_2210 (renews 2026-12-14), scored 0.79 against threshold 0.70, ran q4-renewal-window-v1: note to the account owner with renewal date, current plan, and 4 pages read. 08:41 UTC. Logged.

I matched pricing → plan-compare → repeat (7d) on acct_5138, fit account, scored 0.76 against threshold 0.70, ran q4-budget-check-v1: pages and score by axis to the account owner, account added to the Q4 watch list. 13:07 UTC. Logged.

I matched closed-lost → pricing return (6mo+) on acct_0934 (closed-lost 2026-02-19, reason: budget), scored 0.72 against threshold 0.70, ran boomerang-v1: note to the account owner with loss reason and pages read. Nothing else sent. 16:22 UTC. Logged.

Each of those runs had an owner before it fired, and the owner was not a meeting. The season gets built in September because the signals get owners in September.

Build it this week

Name three Q4 specimens by their parts. Pre-approve one action for each. Put the renewal cohort due before year end on a watch list; every account on it is a Q4 signal waiting to fire. Decide, today, who owns a September signal: a person or a Play. If it is a person, write their name next to the specimen. If nobody's name fits, it is a Play.

Save Your Labor (Day) ends today: first month free on any tier, code at the link.

Stop watching signals. Start running them.