Module 5 — Operating Cadence with KVN · Lesson 5.4
Negation Enforcement
Measuring the value of the things you said no to
~10 min
What you'll learn
- Read the negation-hits surface and interpret it
- Distinguish a silent Negation (decoration) from an effective one
- Use negation-reuse as a workspace-health metric
- Refresh Negations quarterly based on the hit data
Most management frameworks tell you what to do. Few teach you to measure what you said no to. Kavanah's app_kvn_negation_hits table exists for exactly this — to count how often each Negation got cited, in dismissals, in agent decisions, in project deflections. The metric is unromantic and quietly load-bearing. This lesson walks how to read it.
What counts as a Negation hit
Three events produce a negation hit.
A candidate task dismissed citing the Negation. The triage flow has a 'dismiss with reason' option; when the reason references a workspace, project, or persona N, the hit is logged against that N.
An agent decision that cited the Negation. When the AI agent declined to take an action — declined to send an email, declined to spend a budget, declined to make a deal change — and cited a Negation as the reason, the hit is logged.
A candidate redirected. When a candidate task is moved to a different project (or dropped from the workspace entirely) because the Negation said it didn't belong, that's a redirect hit. It is tracked separately because redirects produce different downstream actions than dismissals.
The hit count is what tells you whether the Negation is actually doing work. A Negation that has produced zero hits over a quarter is either irrelevant (the team never encounters the kind of work it's blocking) or invisible (the team is encountering it but the system isn't catching it).
Reading the silence
When a Negation produces zero hits, your move depends on which kind of silence it is.
Irrelevant silence is when the N described a boundary the team never approaches. 'We will not pursue government contracts' produces no hits because nobody is proposing government contracts. The N is technically working — by making the boundary explicit before anyone asked — but it isn't earning ongoing reuse. Keep it; it costs nothing.
Invisible silence is when the team is approaching the boundary but the system isn't catching it. The signs: people grumbling about the kind of work they're doing, projects that look suspiciously like the boundary they were supposed to avoid, candidate tasks getting accepted that on reflection should have been dismissed. Invisible silence is dangerous — the Negation is on paper but not in practice.
Distinguishing the two is straightforward. Ask your team: 'have you encountered anything like X in the last quarter?' For irrelevant silence, the answer is no. For invisible silence, the answer is 'oh yeah, a few times, but we just took them on.' The latter is your signal to either retire the Negation as not real or tighten the triage to use it.
Reading the noise
The opposite of silence is a Negation that produces many hits. This is usually good — it means the boundary is real and the team is bumping against it constantly — but it can also be a signal.
If a single Negation accounts for the majority of dismissals, the question is: should the team's incoming-work mix have been different? A workspace whose top Negation is 'we don't build for sub-10-person teams' and whose dismissal logs are full of sub-10-team requests is telling you something about marketing or sales positioning, not about the rightness of the Negation.
Noisy Negations are also useful for retros. The three-most-cited Negations in a quarter are good material for the team retrospective — they are the boundaries that mattered most, and articulating why they mattered is itself useful institutional learning.
Quarterly Negation refresh
At the quarterly charter revisit (Module 5.1), bring the negation-hits data into the conversation.
For each workspace and project Negation, look at the hit count over the past quarter. Negations with high hit counts get kept and possibly sharpened. Negations with zero hits get the irrelevant-vs-invisible test; irrelevant ones are kept (low-cost insurance), invisible ones are either rewritten or removed.
At the same time, add the Negations that the quarter surfaced. The dismissal reasons that recur across many candidates ('this isn't really our customer,' 'we already said no to this kind of optimization') are candidate Negations that should be promoted into the charter. Promote them; the next quarter's triage gets faster.
Kavanah surfaces the quarterly review of negations in /workspace-kvn and in the project charter — the negation_hits column is queryable and the dashboard surfaces the top and bottom cited Negations. The agent can also produce a draft 'refresh' of the Negation list based on the quarter's data; you finalize.
Run a Negation review this quarter
- 1
Open /workspace-kvn and look at the per-Negation hit counts
Sort by hits. Top of the list is doing real work; bottom of the list needs the irrelevant-vs-invisible test.
- 2
For each zero-hit Negation, ask the team about edge cases
If they recall encountering the boundary and ignoring it, the Negation is invisible. If they have never encountered it, it's irrelevant. Decide accordingly.
- 3
Look at dismissal reasons that recur
Candidates dismissed with reasons that look the same across many cases are candidate Negations. Promote the recurring ones into the workspace or project charter.
- 4
Use the agent to draft a refreshed Negation list
Ask: 'given the dismissal patterns of the last quarter, what Negations should the workspace add or sharpen?' The draft is a starting point.
Negation enforcement
- Total negation hits per quarter
- Sum across workspace, project, and persona Negations.
- Healthy signal: Above zero, usually growing as the team encounters more work. Zero means N's are decoration.
- Per-Negation hit count
- Hits attributable to each individual N over the quarter.
- Healthy signal: Diverse — some Negations heavily cited, some quietly held. All zero or all overwhelmingly cited by one N is a signal something is off.
- Promotion rate
- Number of dismissal-reason patterns promoted into the charter as new Negations per quarter.
- Healthy signal: At least one per quarter for an active workspace. Zero means the charter has stopped learning.
- Negation-induced redirect rate
- Fraction of candidate tasks that get redirected to a different project (rather than dismissed) because of an N.
- Healthy signal: Above zero. Redirects are evidence the Negations are not just stopping work; they're routing it.
Key takeaways
- ·Three events produce a Negation hit: dismissal, agent decision, redirect.
- ·Zero-hit Negations split into irrelevant (keep) and invisible (rewrite or remove).
- ·Noisy Negations may signal pipeline-mix problems upstream.
- ·Quarterly: keep top, test bottom, promote recurring dismissal reasons into the charter.
Operating cadence depends on the charters being load-bearing and the Negations being measured. With those in place, the last module is the full metric framework — what to measure across the whole system to know it is actually working.