Two numbers describe the same season, and they never match. Finance owns the top-down plan: a revenue and margin target set from the top, phased into months, defended in the board deck. Merchandising owns the bottom-up forecast: units by style, size and channel, built from history and instinct. On the first day of the season the two are close enough to sign off. By week three they are describing different businesses.
Everyone treats this as normal. The plan is the plan and the forecast is the forecast, and reconciling them is a quarterly ritual involving a long meeting, a shared drive, and someone exporting a workbook at 11pm. Nobody enjoys it, nobody trusts the result, and by the time the reconciliation is finished the season has moved on again.
The reason the two numbers drift is not that finance and merchandising disagree about the business. It is that they are working in two separate files that were never connected. A number that lives in one spreadsheet cannot know when the number in the other spreadsheet changes. Drift is not a failure of alignment. It is the guaranteed output of the architecture.
Two files cannot stay in sync by meeting
The plan and the forecast drift because nothing connects them but a calendar invite.
Walk the mechanics and the drift becomes obvious. Finance builds the plan in one workbook, phased by month, keyed to dollars and margin rate. Merchandising builds the forecast in another, keyed to units, styles and size curves. The two are reconciled once, at plan sign-off, by hand. Then they are left alone, and the world starts moving.
Sell-through arrives and the forecast updates, or at least the merchant updates it in their heads while the file stays stale. A category runs hot, a launch slips, a supplier is late. Every one of those events changes the bottom-up picture and none of them touch the top-down plan, because the plan has no wiring to the forecast. The only mechanism that keeps them together is a person remembering to update both, and people are busy running a season.
So the gap opens quietly. Finance is still steering to a monthly revenue shape that merchandising abandoned in week two. Merchandising is chasing demand that finance has not re-approved the spend for. Both are working hard, both are right in their own file, and the business is being run off two maps that no longer agree on where the roads are.
Take a concrete case. In January, finance phases twelve million of revenue into the spring months and merchandising builds a unit forecast that ladders up to roughly the same total. Both sign off. In week two, outerwear is pacing ten points ahead and the merchant knows it, so they start planning reorders in their head. But finance is still looking at a plan that assumed the January split, so the working-capital ceiling has not moved and the spend has not been re-approved. The merchant either waits for the next reconcile, which loses the reorder window, or acts and finds out at month-end that they broke a cash constraint nobody had connected to their file. Neither outcome is a mistake by either team. It is the two files behaving exactly as two disconnected files must.
of working capital freed on average when the plan and the forecast run on one reconciled model instead of two files kept in sync by meeting.
That capital does not come from cutting the plan. It comes from ending the drift. When the top-down target and the bottom-up buy are the same object, money stops being trapped between two numbers that disagree, and starts moving to where demand actually is while there is still time to act on it.
The reconciliation ritual is where margin leaks out
By the time the two numbers are squared, the decision window has closed.
The quarterly reconcile feels like control, but look at what it actually is: a backward-looking exercise that tells you the plan and the forecast disagreed, weeks after the disagreement started costing money. You find out in the review that outerwear over-delivered and dresses under-delivered. That is not news you can act on. That is an autopsy.
The cost sits in the lag. A category runs eight points ahead of plan in week two. In a reconciled model you release spend to it and reorder while the season is live. In the two-file world you notice it at the month-end reconcile in week five, by which point the lead time has eaten the reorder window and the moment is gone. The plan and the forecast finally agree, on a number that describes a decision you can no longer make.
The same lag runs the other direction. A category the plan bet on stalls, but the buy is already committed and the reconcile that would have caught it is still three weeks out. The overstock builds while two files quietly disagree about whether there is a problem. By the time they are reconciled, the only lever left is markdown.
There is a subtler cost too, and it is cultural. When the reconcile is a manual quarterly event, it becomes a negotiation rather than a readout. Finance arrives defending the plan, merchandising arrives defending the forecast, and the meeting resolves into a compromise number that neither side fully believes and that describes the business less accurately than either input did. The organization learns that the reconcile is where you argue, not where you see the truth, and the two teams start pre-loading their files with buffer to protect their position going in. The buffers are invisible, they are baked into the plan and the forecast before the meeting, and they are pure over-buy: capital committed to defending a negotiating stance rather than to meeting demand.
Brands on one reconciled model clear more before markdown
The 14-point gap is not better planning instinct. It is the plan and the forecast moving together, in season, instead of being squared up after the fact.
Fourteen points of full-price sell-through is the whole difference between a clean season and a promotional one, and it traces straight back to how fast the two numbers reconcile. Square them once a quarter and you defend a stale shape. Square them continuously and you steer the real one.
One model, two views, always reconciled
Finance keeps its top-down. Merchandising keeps its bottom-up. Both read the same source of truth.
The fix is not to make finance plan in units or force merchandising to think in margin rate. Both views are correct and both need to survive. The fix is to put them on one model so they are two lenses on a single set of numbers, not two files that happen to describe the same season.
On one model, when merchandising's bottom-up forecast moves, the top-down plan sees it the same hour, phased into the months it affects, with the margin and working-capital impact already computed. When finance changes a target, the buy re-plots against the new ceiling inside the guardrails, and the size-and-style detail underneath updates to match. There is no reconcile step because there is nothing to reconcile: the two numbers are the same number, shown two ways.
The guardrails are what let finance step back without losing control. Working-capital ceiling, margin floor, markdown headroom, phasing rules: these are limits the model plans within, not rules a planner has to defend in a meeting. When demand drifts, the agent stages the moves that keep the bottom-up buy inside the top-down plan, and the planner approves the set in one pass instead of rebuilding a workbook.
The plan and the forecast stop being two numbers you argue about and become one number you steer together.
None of this asks finance to trust merchandising's file or merchandising to trust finance's spreadsheet. They both trust the same model, and every change to it is a snapshot in the audit trail: what moved, when, and why. The quarterly reconcile stops being the moment you discover the drift, because there is no drift left to discover.