03 / Decisions & Accountability
Authority Without Consequences Does Not Build Judgment
Giving a manager the authority to decide is not the same as giving them the experience of living with the decision — and only the second one builds judgment the company can rely on.
The example below is illustrative.
The regional manager moves the only technician qualified for a customer inspection onto a late job. The recovery works; the late job ships. Two days later the inspection cannot happen, because no one else can carry it out, and the second customer invokes a delay clause. The manager now has to explain why protecting one customer's date cost the company another's.
On paper this was his 25th recovery decision of the quarter. Of the first 24, most were ordinary scheduling changes made on days when capacity was available. On the genuinely hard ones — the ones that traded one commitment against another — the service vice president had quietly stepped in, renegotiated a deadline, or found a replacement crew. The manager held the authority the whole time. Only now was anyone seeing whether he understood what happens after he uses it.
Inventory the decisions the role has avoided
Authority exercised only on easy cases produces a clean approval log and no evidence about the hard ones. Take one operating decision the manager is expected to own, pull a consecutive sample, and record three things separately: who chose the response, who faced the customer or operating consequence, and who repaired the result. The same person need not do all three — and when they routinely do not, the gap is the point.
| Decision class | What the manager actually carried | Evidence to review |
|---|---|---|
| Swap within an open capacity window | Chose the response and saw the operational result | Schedule change, actual completion, reopened work |
| Protect one promise with a scarce skill | Identified the second customer at risk and chose among feasible alternatives | Qualification records, fallback capacity, customer notifications |
| Recover from a plan that deteriorated | Recognized the change and acted within limits, or escalated in time | Original prediction, new facts, time of intervention, outcome |
A manager can lead the meeting and still have no entry in rows two and three. Where senior help was routine, separate advice from a required approval from a rescue that substituted someone else's decision. Those three are different, and only the last one tells you judgment has not transferred.
Give a decision real but bounded exposure
Start with a recovery that has clear cost and service limits and a fallback if the plan degrades. Give the manager a late job with two feasible routes: a second shift at a known premium, or a re-sequence of two jobs whose dates are still negotiable. Set a written spend ceiling, a named fallback, and a rule that a firm inspection date cannot move without proper approval. Have him verify skill and coverage in the calendar, agree the customer message with the account owner, and track both jobs after choosing.
Price the exposure so the lesson is concrete. Suppose the second shift costs $9,000 and protects a $60,000 order with no downstream risk; the re-sequence costs nothing but has a one-in-three chance of slipping a $40,000 delivery whose contract carries an $8,000 penalty. The expected cost of the "free" option is roughly $2,700 in penalty plus the relationship exposure — cheaper than the shift, but only if the manager can name and accept that risk out loud. The decision he makes matters less than whether he can state the tradeoff before the outcome is known. Arrange a check at the moment the plan can still be reversed, and record any intervention rather than quietly fixing it — a rescue handled off the record gives management false confidence in the result.
Review reasoning with the information available at the time
Before the result arrives, have the manager log what he expects and what would disprove it. Then review the decision and the outcome separately: a sound choice can end badly when an unforeseeable condition changes, and a lucky result can conceal a failure to check an available fact.
A credible progression moves from one-account recovery, to two-account tradeoffs, to cases with incomplete information or a shifting constraint — one new kind of exposure at a time. Stop or narrow the authority when the same missed distinction recurs, and check whether the cause is skill, information, or an incentive that rewards the booking at the expense of the other account. This is the same gap the research traces at the company level in Available Is Not Ready: a resource that looks committable until the specific condition behind the promise is tested. If the expert cannot even explain the distinction the manager missed, that is a prior problem — start by extracting and documenting it. For cases that keep returning upstairs, the returning-decision review separates lack of exposure from missing information or authority.
Take this to your team. Which consequential judgment have we assigned to someone who has never had to see what happens after they make it — and what bounded case could let them feel the consequence while the exposure is still small?
Bring one delegated decision and we will test whether the judgment behind it has actually moved.