Annual Planning While the Quarter Is Off Track Is How July's Numbers Reach the Board

Two tabs, open at the same time. That is what the second week of August looks like for us.

The left one is the Q3 goal sheet. Eleven company priorities, set in the first week of July, and four of them still carry the status they were given at kickoff because nobody has touched them since. One belongs to someone who changed teams in June. Two describe a pricing motion the company stopped running three weeks ago. The sheet is not wrong, exactly. It describes a version of the business that stopped existing somewhere around the third week of July, and everybody can feel that, and nobody has said it in a room where saying it would count.

The right tab is a planning template for next year. Blank except for the section headers and a date: first pass to the CEO by the end of the month. Below it sits a note from the CFO asking for hiring-envelope inputs so finance can start modeling the January headcount plan.

Then the arithmetic nobody says out loud. Whichever tab gets this week, the other one gets worse. Spend the week on the reset and the annual draft arrives thin, and thin drafts get rewritten in October under more pressure and less time. Spend it on the annual draft and the quarter closes on a goal list nobody has believed since July — which happens to be the baseline the annual draft is built from.

We have run this fortnight enough times to stop treating it as bad luck. It arrives every year, in roughly the same week, and the cost of getting the order wrong is not the week itself.

Finish the Quarter First, Even Badly

Finish the quarter's reset first, even a compressed one, before opening the annual plan. A plan built on an unresolved quarter inherits it — assumptions that stopped being true in July get carried into next year's targets, and nobody notices until Q1. The reset does not need to be a replanning cycle. It needs to produce an honest current-state number.

The July Number That Ends Up in the Board Deck

Here is the mechanism, and it is duller than it sounds.

The annual plan opens with a current-state section. Run rate, pipeline coverage, capacity, what actually happened this year against what was committed. Every target downstream is set relative to those numbers. If the current-state section says pipeline coverage is 3.1x, the 2027 revenue plan gets built on 3.1x.

Now suppose that 3.1x was calculated in July, under a pricing motion the company stopped running in the last week of July. The number is not fabricated. It was true when someone typed it. It is simply describing a business that no longer exists, and because it lives in a slide rather than a system, nothing about it looks stale. It reads as a fact.

That number then travels. It sets the sales capacity model, which sets the hiring envelope, which sets the January headcount plan the CFO is already asking for. By December the board has approved a plan whose foundational assumption expired in the third week of July. Nobody catches it, because by then the number has been restated four times in four documents and its origin is gone.

It surfaces in Q1, in the first month the plan meets reality. The conversation that follows is about the Q1 miss. It is almost never about August, because nobody can trace it back that far.

If you take one thing into a meeting this week, take the question that breaks the chain early: which numbers in our current-state section were last pulled before the third week of July, and who would know? It is answerable in an afternoon, and it is the cheapest it will ever be to ask.

Nobody Scheduled the August Collision

Two operating rituals run on independent calendars, and neither calendar was set with the other in mind.

The annual cycle runs backwards from a board date. Growthspree's 2026 B2B SaaS CMO annual planning guide lays the shape out explicitly: five phases across four months, opening with strategic context in August, when the CEO, CRO and CFO give their inputs on next year's revenue plan, hiring plan and capital constraints — then retrospective in early September, hypothesis design through late September and October, budget build and scenario work in November, board approval and rollout in December. It is a marketing-function calendar, and the arithmetic under it is general. If the board approves in December and the plan launches in January, and the design phase alone runs five to six weeks, August is not an aggressive start. It is the last responsible one.

The quarter runs on the fiscal clock, which does not care about any of that. Week six of Q3 is the middle of August. Week six is also, reliably, the week the quarter stops resembling its plan — enough has happened to invalidate the July assumptions, and enough time remains that doing something about it is still worth the effort.

Both calendars are correct in isolation. The intersection belongs to nobody. There is no forum where someone looks at the August calendar and says out loud that one set of people will close one loop and open a much larger one in the same fortnight, out of the same set of hours. That is not a scheduling oversight anyone made. It is what happens when two well-designed cadences are designed separately.

Part of what makes the collision expensive is its timing inside the quarter. A quarter that visibly drifts in week three produces a small correction; the same drift, unnoticed until week six, produces a reset that now competes with annual planning for the same hours. We built Radar around that gap specifically — surfacing the initiative that has stopped moving on day three rather than day thirty, so the August reset is a smaller piece of work than it would otherwise be.

The One-Week Honesty Pass

The reason people skip the reset is that they imagine the wrong thing. They picture a replanning offsite: three days, a facilitator, a new set of objectives. That version genuinely cannot fit in August, so it does not happen, and the quarter finishes on the July sheet.

A reset that fits is much smaller. It is three moves.

1. Establish which goals are still real, and formally kill the ones that are not. Not quietly deprioritize — kill, with a line in the notes saying it was killed and why. Most survive; the ones that do not were already dead in practice, and writing that down converts an awkward silence into a decision somebody made.

2. Reassign anything owned by a person who has moved. Reorgs happen mid-quarter and goal ownership almost never follows them. A goal with a departed owner reads as green until somebody checks, because nobody is there to mark it otherwise.

3. Produce one current-state number per surviving priority that leadership actually agrees on. Not a status color. A number, with the date it was pulled and the source it came from.

Time-box the whole thing to a week. The value of this pass is honesty, not thoroughness, and a compressed honest answer in August beats a rigorous one in October by an enormous margin — because the annual plan is being drafted in the interval between them.

The first two moves are the ones that eat the week, and they eat it in a specific way: not through judgment, but through finding out. Who owns this now. Has anyone updated it. Is this number from before or after the pricing change. That is the work Goals & Alignment was built to remove — the cascade knows who owns what and when it last moved, so the honesty pass starts with the answers rather than with a request for them.

The Gathering Is What Eats You, Not the Deciding

Annual planning has two components that get treated as one, and separating them is most of the fix.

There is the negotiation: how much revenue we can credibly commit to, what we are willing to spend to get it, which bets we are making and which we are declining. That requires senior judgment, disagreement, and time in a room with the CEO and CFO. It is the actual work, and protecting it is the point.

Then there is the input assembly: last year's actuals, current run rate, committed spend, headcount by function, what shipped and what slipped. None of it requires judgment. All of it requires someone to go and get it. And that half is what consumes an operations team for three weeks every autumn.

In a survey of 3,028 middle managers fielded by YouGov over two weeks in early September 2025, 85% said they had proposed a process improvement and only 54% had seen one implemented. The sample was drawn entirely from construction, manufacturing, retail, hospitality and transport rather than software, but that gap is about standing rather than sector, and it describes planning season precisely. The person assembling the inputs can see, in detail, which parts of the assembly are pointless. They are also the person with the least standing to stop doing it in August, because everyone upstream is waiting on the file.

The Pre-Read Brief exists for exactly that split. It pulls from the tools the work already lives in and hands over one page — what moved, what is at risk, what needs a call — so the assembly stops being a person's project and the room can start at the disagreement.

The Ten Seconds It Takes to Check Where a Number Came From

The version of August we are describing is not one where planning gets easier. The negotiation is as hard as it ever was; it should be.

What changes is what has to happen before the negotiation can start. When current state is a standing output rather than a three-week project, the reset stops competing with the annual draft for the same hours. The honesty pass becomes a meeting where people react to numbers already on the screen, instead of a fortnight of asking for them and a meeting where people react to the asking.

The second-order effect matters more and takes a cycle to show up. Reviews that carry context forward mean the August current-state section is not assembled from scratch at all — it accumulates from what the system has already been tracking since April. "Every number traceable to its source" is a plain product claim, and its practical meaning is this: when the 3.1x figure appears in the annual plan, somebody can establish in ten seconds when it was pulled and what it was pulled from. That is a small thing that prevents a large one, which is a decent description of most of this job.

What the Fortnight Is Actually Building

The two tabs are still open. The fiscal calendar and the board calendar were designed separately and will keep colliding, and reconciling cadences nobody owns is a fair description of the job rather than an argument against it.

What changed is the assumption underneath the exhaustion. For a long time we read that fortnight as a personal capacity problem: a better-organized operator would have gotten ahead of it, blocked the calendar, delegated more. That reading is wrong, and it is expensive, because it sends you looking for a discipline fix to a sequencing problem.

McKinsey's analysis of the chief of staff role — drawn from roughly 250 people who have held it across about 300 organizations — puts median tenure at 2.3 years against 4.8 for the average large-company CEO, and finds that two-thirds are promoted afterwards, typically more than one level up. McKinsey does not say why, and we are not going to pretend otherwise. Our own reading is that the years are short because of what they are spent on: fortnights like this one, where somebody holds two operating cycles at once and decides what order they go in. That judgment is what the tenure is buying. The scheduling of it is not, and it should not cost what it currently costs.

Questions About Sequencing Quarterly and Annual Planning

When should we start annual planning for next year?

Phase one — the CEO, CRO and CFO inputs on the revenue plan, hiring envelope and capital constraints — typically starts in August for a January launch, working backwards from December board approval. Starting later compresses the negotiation phase, which is where most of the real decisions get made. Starting earlier tends to produce a plan built on stale current-year assumptions, because the current year has not yet produced enough signal to plan against.

Should we fix the current quarter or start next year's plan first?

Fix the current quarter first, but time-box it hard — a week, not a month. The reason is not tidiness. The annual plan is built on a current-state baseline, and if that baseline is a goal list nobody believes, every number downstream inherits the error and it surfaces in Q1 as a surprise. A compressed, honest reset is worth more to the annual plan than a thorough one delivered in October.

What does a mid-quarter reset actually involve?

Three things. Establish which goals are still real and formally kill the ones that are not. Reassign anything owned by someone who has changed roles or left. Produce a single current-state number per surviving priority that leadership agrees on, with a date and a source attached. It is not a replanning offsite — it is an honesty pass, and it fits in a week if the data is available without chasing it.

How do you keep annual planning from consuming the whole operations team?

Separate the inputs from the negotiation, and treat them as different kinds of work. The inputs — run rate, capacity, committed spend, what actually happened this year — should assemble from systems rather than get gathered by asking people, because the gathering is what consumes the team. The negotiation genuinely requires senior judgment and time in a room, and protecting that time is the entire point of automating the first part.

Request a demo at rhythms.ai.