03 / Decisions & Accountability

Several Correct Numbers, One Unassigned Decision

When four teams report four different true numbers for the same question, the company does not have a data problem — it has an unassigned decision about which number is allowed to govern the promise.

The example below is illustrative.

Sales sees 72 units on hand. Planning sees 40 uncommitted. Quality sees 34 cleared for this customer's specification. Logistics estimates 24 can arrive by Friday. The customer asks for 30 by Friday. All four teams can defend their reports. The company still has to decide what it can promise.

A data cleanup would be justified if the four had measured the same population and disagreed. They did not. On-hand answers a stock question, uncommitted an allocation question, cleared a specification question, deliverable a promise question. The numbers only look like they conflict. Their real disagreement begins the moment someone picks one to authorize a commitment.

Attach a definition to each question

A definition names the population, the time, the requirements, and the exclusions that matter — and shows which restrictions have already been applied. Numbers drawn from overlapping populations cannot be reconciled by taking the smallest total and hoping. Put the customer's item, quantity, location, and deadline beside the definitions, and find the figure that represents the intersection of the necessary conditions.

Prove the number at the level the customer buys

Take the 34 quality-cleared units. Suppose 10 are reserved for another account, six more cannot reach the dispatch site in time, and the remaining 18 have no transport hold. All four original totals stay correct. Eighteen is the most the company can promise under the stated conditions — if a confirmed carrier can move them all. The gap between the tempting 24 and the real 18 is not an arithmetic error; it is the difference between which units Logistics counted and which units satisfy every condition at once.

Put the economics on that gap. If the company promises 30 against the 24 figure and can actually deliver 18, the 12-unit shortfall on a Friday deadline is not just a service miss — if each unit carries $3,000 in contribution and the contract has a short-delivery penalty of $2,000 per unit, promising the wrong number turns a $54,000 order into a roughly $24,000 penalty and a damaged account. The number that governs the promise was worth getting right by $78,000.

Decide who may change the governing condition

The company could split the delivery, pay for a different carrier, or ask another customer to release reserved units — none of which emerges by averaging the counts. Each is a different commitment with its own economics and, often, its own authority. Price the alternatives before changing the rule: a second carrier adds six units only if transportation is the binding constraint; releasing a reservation makes the quantity possible while costing something elsewhere. Give the owner a rule for which definition governs each class of promise, and a clock — establish when the available-to-promise figure expires and who can refresh it while the customer can still use the answer.

Audit promises against the evidence

Take ten recent offers and record the decision, the cited figure, the governing definition, the source evidence, the owner, and the delivered result. Figures that genuinely conflict on the same measure call for reconciliation; a figure built for one purpose and used for another calls for a better decision rule. A defensible quantity can also coexist with an unreleased order — which is a different failure: What Does "Ready" Authorize? covers the state decision that permits someone to act on a promise, where this article covers which quantity may govern it. The distribution mechanism runs the full length of The system shows 47. The customer still can't have 12., and Available Is Not Ready is the synthesis of why "in the system" and "promisable" are never the same claim.

Take this to your team. Pick one promise with several defensible numbers behind it. Which number was authorized to govern the decision — and who could change that answer?

← Back to Decisions & Accountability