The Rebuild Protocol
Detect, name the cost, correct, prove. A sequence, not a sentiment.
When to use this
- A mistake has become visible and the explanation is being drafted before the correction is.
- An apology has already been made and nothing has changed since.
- A team has stopped raising things with you and you know why.
- Credibility was lost in one event and is being rebuilt by general good behaviour, slowly and invisibly.
Trust Retention
The mechanism
Four steps, and the order is the method. Most recovery attempts run them backwards — they prove first, correct later, name the cost never — which is why they read as management of perception rather than repair. Credibility does not return when a mistake is explained well. It returns when people see something change.
- 01
Detect
- What it is
- Establishing that credibility has been damaged, and roughly how much, before anybody tells you.
- How to run it
- Watch the behaviour, not the mood. Are you being copied on more email and told less in person? Have questions become requests for written confirmation? Has the volume of bad news dropped rather than risen?
- What the answer tells you
- Trust withdraws quietly and in a specific pattern: people stop spending informal risk on you and start creating paper. That pattern is your measurement. If you wait for the survey, you are measuring something that finished happening a quarter ago.
- 02
Name the cost
- What it is
- Stating plainly what the mistake cost, in the terms of the people who paid it rather than the terms of the people who made it.
- How to run it
- Say what happened, what it cost, and who absorbed it, in that order and without a subordinate clause defending the intent. Do it once, where the people affected are, and do not do it again.
- What the answer tells you
- This is the step most leaders skip, and skipping it is why the rest fails. A cost that is not named is a cost the organization concludes you did not notice. Naming it accurately is also the only proof available that you understand the problem well enough to be trusted with the fix.
- 03
Correct
- What it is
- Changing the thing that produced the mistake, rather than the description of the mistake.
- How to run it
- Identify the mechanism, not the individual: what made this outcome available? Change it, and make the change visible enough that somebody who was not in the room can see it from the outside.
- What the answer tells you
- If the only thing that changed is how the event is described, the organization has learned that the response to failure is communications. It will route the next failure accordingly, which is to say away from you.
- 04
Prove
- What it is
- Showing the correction working, on a schedule, without being asked.
- How to run it
- Set a date, report against it in public whether or not the news is good, and repeat until nobody is checking any more. Unprompted is the whole point: reporting you were asked for is compliance, reporting you volunteered is evidence.
- What the answer tells you
- Credibility returns on a lag and it returns through repetition, not through eloquence. When people stop asking for the update, the rebuild has finished. That is the only signal worth waiting for.
Personal and enterprise application
A leader runs the sequence in days, not quarters, and usually in one conversation. The failure mode is stopping after step two: the cost is named well, everyone feels heard, and nothing about the mechanism changes.
Detect by watching what stops arriving. Name the cost once. Correct one thing specifically. Then prove it on a date you announced before you knew whether you would make it.
An organization runs the same four steps over weeks, with an owner and a date on each. The sequence does not change with scale — only the number of people who have to see each step for it to count.
The enterprise failure mode is the reverse of the personal one: correction happens, quietly and competently, and is never proved to the people who absorbed the cost. Repair that nobody witnessed does not repair trust.
What good looks like in 90 days
Recovery is falsifiable or it is a feeling. These are the observable outcomes.
- The behaviours you detected in step one have reversed — informal reporting is back and the paper trail has thinned.
- The mechanism you changed has held under at least one comparable event.
- You are still reporting against the correction, and nobody is asking you to.
The Rebuild Protocol
The worksheet
The one-page version, for running the sequence after a visible mistake.