Three questions decide everything.

Every partner described the firm differently to a prospect, so it competed on price against practices it had nothing in common with.

We worked through what ten years of case history actually showed — family-owned businesses, clients staying through generational handovers — and settled one belief the partners would be held to.

Shining Ray holds the belief and every ruling made under it.

A partner turning down work now cites the reason the last one was turned down, instead of relitigating it.

Fourteen dashboards, no two agreeing, and nobody able to say which number the board should act on.

We cut it to four measures that could each name the thing they were evidence for, and retired the rest rather than leaving them up to be argued with.

Compass holds the four, and the lineage under each.

One stopped moving for five weeks and flagged itself — which is how the staffing problem underneath it got found at all.

The same supplier problem every winter, solved from scratch every winter, by whoever happened to be in that morning.

We sat with them through one of them properly — what was blocking, what had been tried, what the constraint really was — and wrote the answer down in the form it would be needed in next time.

Archer holds the heuristic.

The next winter it took an afternoon instead of a fortnight, and not by the same person.

Every roadmap argument restarted from first principles because nobody could point at what the company had already decided it was for.

We separated the two things being argued about — who the product is for, and what to build next quarter — and settled the first before letting anyone near the second.

Shining Ray holds the first; Compass holds the second and shows which belief each bet serves.

Roadmap sessions got shorter, because the half that was really an identity argument stopped happening in them.

Two offices bidding the same kind of work to opposite standards, and no way to tell which one was right.

We walked both through the bids they had won and lost, found the standard the winning ones shared, and wrote it down as something a bid could be checked against.

Shining Ray holds the standard and the rulings made under it.

A junior engineer can now check a bid against the firm instead of against whoever is free that afternoon.

Growth targets that nobody believed, because the measures behind them described the industry rather than this institution.

We replaced them with measures the members would recognise as being about them, and said plainly which ones we could not yet measure at all.

Compass holds them, with the gaps marked as gaps.

The board stopped discussing two numbers that were never evidence of anything, which freed most of a meeting a month.

Depot problems solved brilliantly in one place and never heard about in the other four.

We took the three that recurred most, solved them properly once, and wrote each answer in the form the next depot would need it in rather than as a report about the first.

Archer holds all three, scoped to the situation that triggers them.

The fourth depot to hit the loading problem solved it the same week, from the write-up, without calling anyone.

Every intake brought the same admissions dispute, argued from memory, decided differently each year.

We reconstructed what the department had actually been choosing for, settled it, and recorded the hard cases alongside it as precedent rather than as exceptions.

Shining Ray holds the position and the precedents.

The dispute now takes a morning and ends by citing the year it was last settled.

A retiring floor manager who was, in practice, where thirty years of process knowledge lived.

We spent the notice period turning the parts that were decisions rather than dexterity into written heuristics, with him checking each one.

Archer holds them, filed by the problem that calls for them.

The handover stopped being a cliff. The things only he knew are now things the floor knows.

Funders asked what difference the work made and got a different answer depending on who picked up the phone.

We settled what the charity was actually for — narrower than the mission statement, and truer — then found the two measures that were honest evidence of it.

Shining Ray holds the purpose; Compass holds the two measures and their lineage.

The next funding round was answered with the same sentence by everyone who was asked.