Tool 9 of 32 · an afternoon
Dismissed-Warning Search
Finds the early written warning that already existed, and times how long the company took to hear it.
During Reality, in private conversations, then by locating the dated artifacts those conversations point to.
What it is for
To find whether the correct diagnosis already existed inside the company, early and specific and in writing, and to time how long it took to be heard.
The failure it catches
The company does not have a knowledge problem. It has an admissibility problem. The warning was already there, and it was dismissed.
Why it works
Only a written, dated artifact can beat hindsight. The gap between first raised in writing and acted on is a number about how warnings travel, which more analysis cannot improve.
How to run it
- In private conversations, ask two questions. Who saw this problem earliest, and where is that written down. And what is harder than it should be, that you have been saying so about for a while.
- Find the written artifacts. Verbal-only warnings get noted but do not enter the timing. Their absence from writing is a separate finding about what this company lets people write down.
- For each artifact, mark two dates: first raised in writing, and acted on, meaning behavior changed, evidenced by a decision, a re-plan, a commit, or a reassignment. Agreement in a meeting without a behavior change is not action. If never acted on, the second date is today.
- Time-to-vindication is the gap. Report the set, not just the worst case.
- Deliver the finding before naming its author. Truth first, credit second.
How to read the result
Measured in weeks, information paths are slow but alive. Measured in months or years, with the caller dismissed, the company's real problem is not analysis, and more analysis will not help. What needs to change is the path a concern travels and what happens when it arrives.
How to run it well
- Hindsight makes everyone a prophet, so only written, dated records count.
- This tool is about fixing the path, never about crediting or blaming individuals.
Where it breaks
- Hindsight makes everyone a prophet, so only written, dated records count.
- If you run it as a hunt for who was right, you will teach people to stop writing things down, which makes the company sicker.
- Never surface the author to leadership before the finding has landed as true.
What would prove it wrong
A company with long historical time-to-vindication that nonetheless acts fast on new warnings would separate past from current admissibility. Re-test on a live warning.
What we watch
Whether long time-to-vindication still predicts that a live warning raised this month will stall, or whether the company has already learned to hear new ones.
Ask yourself
- Who called the current problem earliest, and where is that written down?
- What is harder than it should be, that someone has been saying so about for a while?
- How long sat between the concern first being written and the company changing its behavior?
- Was the warning raised more than once, and did the way it was said get softer each time?
- What would have had to be different for that warning to land?
Feeds: Baseline Vitals