The apparel planning canon is built on a rhythm: commit part of the buy, read the season, then act. Chase the winner, reallocate away from the loser, hold an open-to-buy reserve for the things nobody predicted. It is a good rhythm, and it has produced most of the vocabulary the industry uses. It also quietly encodes an assumption — that there will be a second decision — and that assumption is doing more work than it appears to.
Several categories operate without it. A toy line is committed against tooling, licensor approval and ocean transit long before anyone has a read; by the time the peak arrives, the only levers left are where the inventory sits and when it clears. A juvenile hard-goods platform is engineered and certified years ahead, and a chassis decision cannot be revisited when the read comes in. A gifting-led jewelry season concentrates into windows that are over before a reorder could physically land. These are not categories that plan badly. They are categories where the plan is very close to the whole outcome.
The correction opportunity is a hidden subsidy
What a chase actually provides is forgiveness. It absorbs a size curve built from receipts instead of stockout-corrected selling. It rescues a depth decision that was made on instinct. It compensates for a demand signal that nobody put in the plan. In a category with a real chase capability, a moderate amount of upstream sloppiness is survivable, and so it survives — often for years, unexamined, because the outcome kept landing acceptably.
Remove the chase and every one of those upstream shortcuts becomes visible at full cost. This is why teams moving from apparel into an adjacent category are so often surprised by how exposed they feel: the discipline they brought was calibrated against a safety net that is no longer under them. The work does not change. The tolerance does.
Three things carry the weight instead
When there is no second decision, three upstream capabilities stop being good practice and start being the whole defence.
The evidence base behind the first decision. A curve built from what shipped rather than from what sold is wrong in a specific, systematic direction — it under-reads whatever ran out. In a chase category that error gets corrected by the chase. In a no-chase category it is baked into the whole season. Stockout-corrected sell-through stops being a refinement and becomes the difference between a plan and a guess.
Any genuinely forward-looking signal. Most demand signals are projections of the past, which is why they are weakest exactly when they matter most. A few categories have something better and routinely fail to use it — a baby registry is created months before it converts, itemised and explicit about intent; a retailer’s committed order is a real forward commitment backed by their own planning. In a chase category, leaving those in a channel report is a missed opportunity. Here it is the difference between planning forward and extrapolating backward.
The decisions that close before anyone notices. Pack configuration is fixed at packaging design. A chassis is fixed at certification. Metal content is fixed at specification. Each of those closes a door on the plan’s flexibility, and each is typically decided by a function that never sees the open-to-buy consequence. In a chase category the downstream team absorbs it. Without a chase, an upstream decision made in isolation simply becomes the outcome.
What this asks of the system underneath
All three of those are, structurally, the same request: the constraint and the decision have to be in the same place at the same time. A pack size that lives in a packaging file cannot inform a depth decision made in an assortment spreadsheet. A registry signal that lives in a channel report cannot inform a buy committed on a different cadence by a different team. A metal cost that lives with finance cannot inform a mix decision made by merchandising in month two, when respecifying is still available.
This is the disconnected-stack problem described in the pattern beyond apparel, arriving with the tolerance removed. In apparel a fragmented stack produces margin leakage that in-season trading partly recovers, and the recovery hides how much was lost. In a no-chase category the same fragmentation produces the outcome directly, with nothing downstream to soften it. The categories with the least room for error are, awkwardly, often the ones running the most fragmented tooling — because they are smaller, more specialised, and less well served by the software market.
The related observation from the margin is made before the buy generalises here and sharpens: if margin in apparel is largely constructed before the season, then in a no-chase category it is constructed and concluded before the season. The window in which a decision can still change the outcome is not merely early. It is the only window there is.
The uncomfortable implication
Categories without a chase tend to be described — including by the people in them — as inherently riskier than apparel. That framing is not quite right, and it is unhelpful, because it treats the exposure as weather rather than as something a system produces.
The risk is not that the demand is less knowable. Toy demand is highly seasonal but not especially mysterious; a registry is a better forward signal than almost anything apparel has; a chassis lifecycle is known years in advance. The risk is that everything knowable has to be brought together at one moment, and most organisations are not organised to do that. The information exists. It is distributed across functions that meet on different cadences, and the plan is committed before they converge.
That is a solvable problem, and it is the one worth working on. Not better forecasting — better convergence.