Planning Cycle

Setting goals happens at every company. What rarely happens is taking the goals back out and looking at them. Five things decided in January are forgotten by March, and the work that actually got done in between is whatever looked urgent that week.

Without a fixed cycle, decisions follow events. A large customer asks for one feature and the development plan changes that week; a competitor ships something and it changes again the next. Each move is reasonable, and put together they point nowhere. Six months later what's left is five or six things that were started and never finished.

The worse part is that conflicts have nowhere to surface. Pagedeck went through a quarter like that. Mia had told an enterprise account that the permissions feature they asked for would land inside the quarter, and over the same three months Theo's engineering team planned to spend two months on deferred system cleanup. Both plans were right where they stood. The problem was that both needed the same four people, and nobody put that in front of anyone. It came out halfway through the quarter, when Mia asked about the schedule, and by then the customer already had a date.

At a company with no cycle, the first place this shows up is usually a number. Ten hires were budgeted at the start of the year, six have landed by the third quarter, and no one can explain the four that didn't. Everyone knows the situation in their own team. There's just no seat where the whole thing is visible.

A planning cycle is the frame built to handle both of those at once. You decide what to do at the start of a period, force overlapping people and money to collide right there, check progress partway through, and set the next period against the result. What makes it work is not planning well but repeating the same motion at a fixed interval. That's also why this group sits under operations inside strategy, its parent field. Choosing what to do belongs to strategy; loading that choice onto a period and turning it is what the planning cycle does. If corporate strategy is the bundle of choices about which customers to serve and which strength to win with, the planning cycle is the procedure that converts those choices into three-month or twelve-month goals and staffing. A cycle without a strategy turns into a quarterly meeting for triaging urgent work, and a strategy without a cycle lives only inside a deck.

The terms under the planning cycle fall into four groups: how often it turns (cadence), what gets decided when it does (decisions), how progress is watched while it runs (review), and how the results reach everyone (communication). The four run in time order. Cadence creates the occasion, decisions get made at it, review attaches while the period runs, and communication spreads the outcome.

Cadence splits the interval at which planning starts over. There are two, divided by how long a decision holds once it's made.

1. Annual plan — the goals and budget for a year. How much revenue to aim for, how far headcount grows, how much goes to marketing. It isn't a document trying to predict the next twelve months precisely; it's a document deciding where the large resources are committed. Salary increases and hiring numbers are bound to the year, so they have to conclude here.
2. Quarterly planning — resetting goals and priorities every three months. It cuts the direction the annual plan pointed at into pieces the quarter can hold, and folds in whatever changed since the last one. The difference is that the annual plan is not the place to change direction, while quarterly planning is the place to adjust it.

The two intervals differ for a reason. Redecide everything monthly and planning eats the time meant for the work; decide once a year and a situation that changed in June waits until the following January. Most companies run annual and quarterly on top of each other to avoid both failures at once. Pagedeck locks next year's annual plan in December and re-picks the work inside it each quarter. It sets a marketing budget of 600 million won at the start of the year, then decides each quarter how much of that goes to which channel. Big numbers move once a year and small decisions move every three months.

The length of the cycle depends on how fast the company changes. A company whose product hasn't found its footing may find a quarter long and run on a month or six weeks; a company selling to enterprises on long contracts often fits a half-year better. There's no correct answer, and the test is finding the point where the time spent planning doesn't stand out against the time spent working in that period.

Decisions splits the judgments actually made at the planning table, and there are two.

1. Resource allocation — deciding how much of the people and budget goes where. It's the first collision in any plan, and what it exposes is that the sum of what everyone wants to do is always larger than what's available. Splitting twelve engineers across three initiatives means nothing outside those three gets done. If the plan lists eight items and there are only enough people for three, the other five aren't things you decided against; they're things you'll pretend to do and fail to finish.
2. Company priority — ordering initiatives across different organizations at the company level. It behaves differently from ordering inside a team. Within one team you can compare which item is more urgent, but a feature sales asked for and a cleanup engineering deferred are judged by different people. The only seat that can rank them is the one above both. Leave that seat empty and the order gets set by whichever team is loudest or whichever customer complained most recently.

The two are really the front and back of one decision. Ranking gives resource allocation its justification, and running short of resources is what forces the ranking. This is also the stretch where planning meetings run long.

Review covers watching progress while the period turns, and there are two — one routine, one exceptional.

1. Business review — a meeting that checks results and metrics at a set interval. It runs monthly or once a quarter, puts the planned numbers next to the actual ones, and works out why they differ. It isn't a venue for reporting what went well but for catching what went off early. If an item sits at 60 percent of target, what changes for the rest of the period gets decided right there. When the meeting ends at reporting, the same number gets read aloud for three months and the quarter passes.
2. War room — pulling the relevant people into one place temporarily for a single urgent problem. You reach for it when the routine review can't move fast enough: a major account about to churn, a severe outage, anything needing a judgment call daily. Temporary is the whole of its character, so opening one without an exit condition turns it into a standing team that eats the original plan.

Communication carries what was decided to everyone, and it holds one seat. An all-hands is the meeting where the whole company hears the current state together. What was chosen for this quarter, what landed last quarter and what didn't, in the same room and the same words. Without it the plan is shared only among the people who wrote it, and everyone else spends the quarter not knowing which company goal their work hangs from. Whether the misses get named is what decides the value of this meeting. An all-hands that reports only wins stops being believed after two or three rounds. Of forty-five people at Pagedeck, six or seven sit in the planning meetings; for the other thirty-eight this is the only channel through which company decisions arrive.

"That isn't in this quarter's plan. If it goes in, we decide together what comes out."

Not a refusal of the request — the resources are fixed, so it can only be taken as a trade.

"The quarterly review came in at 60 percent of target. Let's settle next week what changes in the remaining two months."

Refusing to let the number end as a report, and carrying it into decisions for the rest of the period.

"Let's put this one in a war room, on the condition that we're back to normal operations in two weeks."

The regular cycle responds too slowly, so use the temporary setup but nail down when it closes.

Pagedeck turns on a quarterly cycle. Two weeks before the quarter starts, Grace collects each team's draft, and Mia, Theo, and Ravi put what they intend to do on the same table. Overlaps surface the moment names appear on that table. Last quarter it came out there that three teams had built their plans around the same two backend engineers, and it was settled by pushing one of the three to the next quarter.

Midway through the quarter, Grace and the team leads meet once a month to look at the numbers. When the second month's meeting showed new signups at half of target, Ravi dropped the large event he had been preparing and moved that budget to search ads. The week after the quarter closes, the results go to the all-hands. Grace gives the two goals that were met and the one that wasn't the same amount of airtime, and for the miss she says whether it gets another attempt next quarter or gets dropped.

The most common belief is that doing this well means sticking to the plan as written. A plan reads like a promise, so that's a fair assumption. But pushing through what was decided three months ago requires ignoring everything learned since. A team that keeps building the original item while twenty customers spend the second month describing a different problem finishes the quarter with 100 percent plan attainment and one feature nobody uses. Changing the plan isn't the failure; letting the change go unrecorded is. When you change it, capture what changed and why, in the same room.

Treating the cycle as paperwork is the other frequent mistake. Filling in a template and entering numbers is the visible part, so it feels that way. What the procedure actually produces, though, is not a document but an occasion to sit across from each other. Three teams planning around the same person surfaces because three teams look at one table in one room, not because the table was formatted nicely. Have each team fill in the template and email it, and the documents exist while the conflict stays exactly where it was. Quarters like that return the same problem at a higher price around the halfway mark. What decides whether this procedure works is less the polish of the document than whether the people holding the decisions are sitting there.