The words we use, and what we mean by them.

The four-week diagnostic every engagement opens with: where the money, the hours and the risk go, measured against the client's own data, and a costed plan. Its first recommendation can be that the client does not need us.

A record that lets a decision be reconstructed later from the data that produced it, who, what, when, on which inputs. The thing every examiner asks for.

The published fee range for a service or the audit. Quoted fixed before the work starts; a band that cannot be quoted is a scope we have not understood yet.

The value of the outcome figure before the work, measured the same way it will be measured after. Without one, a result is a claim.

The call on whether to build a capability or purchase it. Ours is evidence rather than a commission, because we resell nothing.

A team able to run and improve what was delivered without us. The People & Organisation service is measured on it.

Revenue less the costs that vary with it, the margin that tells you whether a product, store or customer is paying for itself after the costs it actually causes.

A system that proposes an action to a person and shows the evidence for it; the person decides. Scoped to one job each.

Ownership, definitions, quality gates and lineage, the eight weeks that decide whether a model holds. Nobody puts them in the press release.

Who decides what, written down. An operating model without them is a diagram of a team.

Weeks 5–12 of the method: the answer on paper, tested in one site, queue or segment before it is built, and the outcome agreed in writing.

One piece of work with one client, closed on one measured outcome. Sixty-four since 2021.

The number a system must reach before launch and hold afterwards, agreed in writing with a named owner. Argue it in week one or you will argue it in week forty.

A price for a scope, quoted before the work starts and not metered. Our only fee model, in four shapes.

Weeks 25–28: the operating model, the runbooks, and the rehearsal, twice, the second time with us silent. Then the last invoice.

The map from a source field to the figure it ends up in, every transformation shown. The artefact that pays for the rest of a data foundation.

Contribution rebuilt from the ledger by product, site and basket rather than by the categories in the management pack. Usually finds a fifth of something paying for the rest.

The four stages every engagement follows (audit, design, delivery, handover) twenty-eight weeks end to end, each stage bought on its own.

The section on every service page that says what the service is not. A page that only sells is a brochure.

The outcome of an engagement the client's agreement does not allow us to print. Fifty-two of sixty-four. We will talk about them.

Named roles, their decision rights and the rhythm they work to. Not an org chart.

The audit's main output: every candidate move, priced, ranked, with the ones we advise against marked.

The one figure an engagement is judged on (with its unit, baseline and date) agreed before the work and published either way.

A trial designed to answer 'can this work?' under conditions chosen to make it work. We argue against them.

One plan, one risk register, one report, and a director who can stop a workstream. Ours is handed to the client by month six.

An outcome the client has approved in writing for the site, with how it was measured. Twelve of sixty-four; two smaller than hoped.

The weekly or monthly value of an improvement once it is live, tracked against the baseline. A saving that exists only in a business case is not a saving.

What the client's team does on the Monday after we leave, written by them with us reading over their shoulder.

One of nine we work in; the counts are derived from the ledger and shown on the sectors page.

The share of delivered change still running a year after handover, 91% of the 43 engagements old enough to check. The firm's own scoreboard.

One job, end to end, in production, with a named owner, what we build instead of a pilot.

The rollout rule: the queue, store or route with the worst service goes first, so the hardest cases prove the change rather than the easiest flattering it.