Terminal stock, and the intake that made it

The units still there when the window shuts

When a selling season ends there is usually a pile of unsold garments in the warehouse. Everybody can see it. This lesson prices that pile at what it will honestly fetch. It then splits it into its only two causes: too much ordered, or not enough sold.

The other half of the problem leaves nothing behind to point at. A line that ran out early sent customers away, and it appears in no report anywhere.

Terminal stock is what is left when the selling window closes. It is not "slow stock" or "aged stock". Those describe something in season that you can still fix. Terminal stock is the residue: the units that reached the end of the period the business planned to sell them in.

MWD-2260's full-price window shut on Sunday 21 June 2026, at the end of week 20. There were 4,732 units left. Here is what became of them.

Those 4,732 units cost GBP 22,335 to land. They brought back GBP 24,330. So the whole terminal block — six weeks of the summer sale, a third of the shop's jersey space, every markdown ticket, the consolidation and the baling — produced GBP 1,995 of gross margin. That is a margin rate of 8.2% on a style whose intake margin was 64.6%.

Sold at full price, the same 4,732 units would have made GBP 40,758. So the terminal stock cost Marchwood GBP 38,763 of gross margin, against a season in which they had all sold.

The identity that produces terminal stock, and its two halves

Step up from one style to the whole class. Marchwood plans jersey tops in units, and values the stock at cost. That is worth saying once, because a plan kept at retail selling value answers a different question and will disagree with this one about the same season.

The intake a plan needs comes straight out of the stock identity, rearranged:

intake required = planned closing stock + planned sales − opening stock

That is the same bookkeeping as lesson 1, read the other way round. It is also the units skeleton of open-to-buy, which is the same idea in money, with orders already placed subtracted from it. Building and trading an open-to-buy is course 17.2. This is the arithmetic underneath it.

Check that both columns close before you read anything into them. Plan: 6,200 plus 96,000 less 7,400 is 94,800. Actual: 7,400 plus 101,300 less 94,100 is 14,600. Both identities hold. That means the argument that follows is about decisions rather than about bookkeeping.

Marchwood ended the season with 8,400 units more than it planned to own. There are only two places that can come from.

That is not an approximation, and it is not an allocation. It is arithmetic. Excess closing stock is over-intake plus under-sales, always, because those are the only two terms in the identity that can move.

The split matters because the two halves belong to different people and have different remedies. Under-sales is a trading problem: the customer did not turn up, or the product was wrong, or the weather. Over-intake is a buying problem, decided months earlier. It is the larger half here — 77% of the excess. Of the 6,500 units of over-intake, 4,500 are MWD-2260 alone, which was bought 18,000 against a plan of 13,500.

At a class average landed cost of GBP 5.20, those 8,400 units are GBP 43,680 of Marchwood's money sitting in stock nobody meant to own. At the class average ticket of GBP 18.50, they are GBP 129,500 of retail value that has to be disposed of somehow.

The half of the story that leaves no stock behind

Everything above is countable, because it is objects. Now the other style.

MWD-2214 sold 8,882 units at full price and finished with nothing. Its full-price sell-through was 92.5%. Its achieved margin was 65.8% against an intake margin of 67.0%. By every number on the sheet it was the best thing Marchwood bought all spring. It also ran out of stock in week 14 of a twenty-week window.

How much demand met an empty rail? That figure exists nowhere, so it has to be estimated, and the estimate has to be defensible.

Marchwood's method is to take the last week the style was selling out of a complete size run, then apply the class's own observed rate of decline. Week 11 is the last clean week: 660 units. For the rate of decline there are two in-house readings. MWD-2260 fell from 784 in week 4 to 580 in week 14, which is 3.0% a week. MWD-2214's own weeks 8 to 11 give 2.6% a week. Use the faster one. It produces the smaller answer, and this is an argument you will have to defend.

Add the 7,762 units actually sold in weeks 1 to 11. The modelled full-price demand across the whole window is 12,878 units. Marchwood bought 9,600. So 3,278 units of demand met no stock — 34% more than was bought.

At a full-price gross margin of GBP 12.28 a unit, that is GBP 40,265 of gross margin that never happened.

The fuller comparison is worse, because a deeper buy also avoids the units that had to be discounted at the end.

GBP 46,264 of gross margin, on the line that looked like the season's success story. It is larger than the whole terminal-stock problem on MWD-2260. And unlike that one, it left no object anywhere in the business for anyone to point at.

The week it was all decided

Both of these outcomes — 8,400 units of stock nobody wanted, and 3,278 units of demand nobody could serve — were fixed before the season opened.

Marchwood's lead time with Aviyur Knits is sixteen weeks: eleven to make and five to ship and clear. The spring 2026 quantities were committed on Monday 13 October 2025. The first sales evidence arrived in week 1, on 2 February 2026. By week 8, when the sheet said everything this course has said, a fresh order would have landed in week 24 — three weeks after the full-price window shut.

So the honest description of the season is this: the two biggest numbers in it were decided sixteen weeks before there was any evidence, and no in-season number could reverse either. That is not a failure of the weekly sheet. It is the shape of the job.

What the weekly sheet can do is worth being precise about, because "it was decided in October" is an excuse if you stop there.

1. It can change what a markdown costs. MWD-2260's cover said 23.5 weeks against a 17-week window in week 3. A markdown taken at week 10 rather than week 15 sells more units at 25% off, and fewer at 70% off and to a jobber. Pricing that trade properly is course 17.4, and it is worth real money.
2. It can move units. MWD-2214's last 718 units went at 30% off and to a jobber largely because they were scattered across 54 shops in the wrong sizes. Consolidated at week 12 into the shops still selling them, a good part of that would have gone at GBP 22.00.
3. It can correct next October. This is the one nobody does, and it is the largest.

The number that has to be written down, or the season repeats

Next year's buy for the striped tee will be built from this year's history. This year's history says MWD-2214 sold 9,600 units.

It did not. It sold 9,600 units because that is how many there were. The sales record of a style that ran out is a record of the stock, not of the demand. Unless somebody writes 12,878 next to it, with the method and the date, the file will quietly say 9,600 forever. The buy will be repeated at 9,600, and the same 3,278 units of demand will go unserved every spring until the style is dropped for underperforming.

That is why the estimate in this lesson had to be made at all, and why it had to be conservative enough to survive an argument. It is not an exercise in blame. It is the only mechanism by which a business finds out how big its own demand is.

Act as a retail merchandiser closing a season down. For the class or department I name, my PLAN and my ACTUAL for each of: opening stock units, sales units, intake units, closing stock units. Also: my class average landed cost a unit [AMOUNT], my class average ticket [AMOUNT], and the styles that make up the largest part of the closing stock, with the units bought against the units planned for each. Do the following. First, check that both columns close on the identity — closing stock equals opening plus intake less sales. If either does not, stop and tell me which, because the whole argument below is meaningless on a sheet that does not balance. Second, give me the excess closing stock against plan and split it into exactly two parts, over-intake and under-sales, with the arithmetic shown. State plainly that these are the only two terms in the identity that can move, so the split is arithmetic and not an allocation. Third, express each half as a share of the excess, and name which is larger. Fourth, value the excess at cost and at retail, and say which of the two figures answers which question. Fifth, attribute the over-intake half down to individual styles, so the conversation is about a decision somebody took rather than about a department. Sixth, for each of those styles, tell me the week in the season at which its cover first exceeded the weeks of window left. That is the week the outcome became visible, and it is the answer to 'when could we have known'. Seventh, ask me whether any style in the class SOLD OUT before the window closed. If so, tell me plainly that its excess is negative and invisible in this table, and that estimating it is a separate piece of work. Do not offer remedies for the under-sales half until the split is agreed, and never report the closing stock as one number.

What you can do now

You can do the arithmetic this track is built on, on real figures, without waving at any of it.

1. Read a weekly sheet: check that opening plus intake less sales plus returns closes to the stock on hand, and refuse to interpret one that does not.
2. Work out a rate of sale over a stated window, at chain level and per shop per week.
3. Quote a sell-through with its week attached, gross and net of returns, and say which you are quoting.
4. Calculate forward cover and read it against the weeks that are left rather than against a feeling.
5. Convert markup to margin and back, and build a ticket price from a target margin and a tax rate.
6. Separate intake margin from achieved margin, and price the distance between them.
7. State the full-price sell-through a plan requires, together with the disposal assumption it rests on.
8. Price terminal stock at what it actually fetches, and split the excess into over-intake and under-sales.
9. Estimate the demand a sold-out style never got to meet, conservatively, and write it into the record.

And one habit that is not a calculation. Every number in this course was available to Marchwood before the outcome was, and most of them were available in week 8. The arithmetic is not the hard part of retail planning. The hard part is doing it early enough that it is still a decision, and writing down what it said, so that next October's buy can read it.