A signal or a hope: what would it cost them to change their mind?

One question, asked of everything

A factory hears dozens of statements about the future every season. A few are worth spending money on. Most are not. The difference is not honesty. Almost every one of them is honestly meant.

One question separates them. Marrash now writes the answer next to every statement in its ledger:

What would it cost this person to change their mind?

That is the whole test. Not how confident they sounded. Not how senior they are. Not how long you have worked together. What it would cost them, personally, to take the statement back tomorrow.

The ladder, priced from Marrash's own book

Marrash went back through six seasons. It sorted every statement it had received by what the buyer had already given up when they made it. Then it counted how often each kind turned into shipped goods.

The counts behind those rates are 6 of 23, 7 of 19, 26 of 41, 14 of 19 and 21 of 22. They are small numbers. Small is what a real factory has.

The ladder climbs as the cost climbs. That is what you would expect. The interesting part is the top three rows. Those three cost the buyer no money at all, and they converted at 26.1%, 36.8% and 63.4%. So money alone does not put them in order.

What puts them in order is who carried the cost. A verbal indication costs the person saying it nothing. A written forecast costs them a template. A capacity enquiry costs a planner a slot in their own process. It also costs them credibility inside their own building, if they ask three vendors to hold weeks and then place nothing.

So the rule needs one refinement, and this is the useful half: the cost that counts is the cost to the person who made the statement, not the cost to their company. A buying director's casual encouragement costs the company nothing and costs them nothing. A junior planner's capacity enquiry costs the company nothing and costs them something real.

Act as a sceptical factory owner who has been wrong about a buyer before. I am going to give you something a buyer told me. Score it, do not interpret it. The statement: [PASTE IT VERBATIM, INCLUDING WHO SAID IT, THEIR ROLE, THE CHANNEL - CALL, EMAIL, MEETING, MESSAGE - AND THE DATE]. Context: buyer [NAME], our history [NUMBER] seasons, [NUMBER] styles shipped, [NUMBER] repeated. What they have already committed on this programme, if anything: [SAMPLES ORDERED AND PAID, LAB DIPS APPROVED, FABRIC AUTHORISATION SIGNED, TECH PACKS RECEIVED, PRICES AGREED, PURCHASE ORDER - AND THE DATE OF EACH]. What I am thinking of committing: [CAPACITY, YARN, FABRIC, TRIMS - WITH QUANTITY, MONEY AND THE DATE IT BECOMES IRREVERSIBLE]. Do the following. First, restate the statement as a fact about what the buyer DID, with the interpretation stripped out, in one sentence I could put in a log. Second, answer the only question that matters. What would it cost this person, personally, to withdraw this tomorrow, and what would it cost their company? Say so if the answer is nothing. Third, place it on a commitment ladder that runs from a verbal indication up to a purchase order. Tell me what the next rung up would be and exactly how I would ask for it. Fourth, name the level the statement was made at - their total, their category, their vendor split, my book - and warn me if I am about to read a number about their total as a number about my share. Fifth, tell me what I would be committing beyond what they have committed, in money and in days. Say whether that gap is one I should buy on purpose or refuse. Sixth, write out the one message I could send today that would move this up a rung. Never tell me a statement is reliable because the relationship is long. Tell me what has been spent.

The accurate forecast that was wrong

Now the case that gives the lesson its title. Nobody lied anywhere in it.

Ovenden gave all its vendors a seasonal forecast: 214,000 pieces for the year. Ovenden actually bought 209,000. So the forecast was 97.7% accurate. That is better than most companies manage about themselves.

Ovenden's brief implied that 96,000 of those would come to Marrash. Marrash received 61,000. The other 148,000 went elsewhere. And 61,000 plus 148,000 is 209,000, so the total held perfectly.

Marrash got 63.5% of its implied share. It was short by 36.5%, against a forecast that was 97.7% right.

Marrash had committed against the 96,000. It had booked greige — undyed, unfinished cloth, straight off the loom or the knitting machine — for 22,000 pieces of the gap. That is 9,240 kg at USD 4.05, so USD 37,422 of cloth sitting in a warehouse. Holding it for 19 weeks cost USD 1,893.55. Moving it into the factory's own basics programme cost USD 2,956.80 in discount. That is USD 4,850.35 on one misread number.

The figure is small on purpose. This is what the mistake costs when the factory has committed only a little. Lesson 6 is about keeping it that way on purpose, rather than by luck.

What to write in the ledger

Restate every statement before you file it. Not Ovenden expects a big autumn — that is an interpretation, and you cannot score it. Write it like this instead:

• Ovenden's merchandiser said on a call that autumn looks strong. Cost to take it back: nothing. Base rate 26.1%.
• Ovenden asked for a price on 48,000, then again on 72,000. Cost to take it back: no money, and a costing team's time on both sides.
• Ovenden signed a fabric authorisation for 20,160 kg. Cost to take it back: the fabric.

Three rows, three very different numbers. The third one is the only one Marrash would buy dyed cloth against.

And here is the rule the whole ladder gives you. It is worth more than any single rate: never commit more than the buyer has committed, unless you have priced the difference and decided to buy it on purpose. Going further is sometimes right. Going further without noticing is what the ladder is there to stop.