Tool 7 of 32 · a few minutes
Padding Probe
Two private questions that reveal whether estimates are still information or have become negotiation.
During Reality, as two private conversations, staff side first, leadership side at least a day later, never in a meeting called for this purpose.
What it is for
To find out whether estimates are still information or have become a negotiation both sides are playing, by asking each end of the conversation the same thing, in private.
The failure it catches
Staff pad because leaders cut. Leaders cut because staff pad. Every date becomes negotiated fiction and blow-ups feel sudden, even though the information existed in someone's head the whole time.
Why it works
Each side will describe the loop from their own end if they are not performing for the other. Asked in private and compared, the two answers either name the same arms race or they do not. A group cannot produce that reading.
How to run it
- Ask the people doing the work, in private: what happens to your estimates after you give them.
- Ask the people receiving estimates, in private, at least a day later: what do you do with the estimates you receive.
- Do not call a meeting for this. Ride inside conversations that were already happening.
- Compare the two answers. The finding is a number about the company, not a quote about a person.
- Never quote either side's answer to the other with a name attached.
How to read the result
If both sides independently describe the loop from their own end, estimates are not producing information anymore. Every number leadership sees is managed. The fix is a process fix: estimates carry their assumptions with them, and revisions happen in the open, so accuracy stops being punished. One cautious estimator is normal professional care, not a broken process. If one side describes the loop and the other denies it, that asymmetry is its own finding.
How to run it well
- Genuine privacy is the whole tool. No one else present, nothing in a group channel, nothing in writing that can be forwarded.
- Two or three people per side is a reading. One person per side is a sample.
Where it breaks
- This only works with genuine privacy. Broken privacy contaminates the answers.
- Never quote either side's answer to the other with a name attached.
- One cautious estimator is normal professional care, not a broken process.
What would prove it wrong
A company where both halves describe the arms race but delivery dates are reliably met would break the probe's meaning.
What we watch
Whether both-sides descriptions of the loop later match missed dates and managed numbers, or whether some companies negotiate estimates and still land.
Ask yourself
- What happens to estimates here after they are given?
- What do the people who receive estimates actually do with them?
- Between the number in someone's head and the number they say out loud, is there a difference, and what drives it?
- When a date blows up, did anyone already know, privately, that it would?
Feeds: Forecast Ledger