Bring the unfinished thought

Northstar is 38% of ARR, the company has eight weeks of runway and the buyer wants a dashboard that could expose private coaching topics. Every related Socra starts from the same evidence and decision rules.

What this Journey remembers

• Northstar requests department-level coaching topics and at-risk labels for its 600-seat pilot.
• Nineteen of 28 employees expect to withhold details. Matched topic-report and program-report usage is ready to compare.
• Departments have four to seven active participants. The privacy model requires ten contributors, caps one person's share at 20% and tests new releases against earlier reports.

Journey instruction

When the renewal decision is resolved, preserve the approved decision, reasons and privacy limits.

Socra makes a call

Northstar says topic access is required to renew. Participant interviews and usage data test whether that access earns enough buyer support to justify lost participation and privacy risk.

1. Maya Chen

   @socra, Northstar says no manager dashboard, no renewal. Do we accept their terms, counter, or walk away from 38% of ARR?

2. @socra

   Counter with a Coaching Program Report that shows whether coaching works and prevents managers from reconstructing a conversation. The deciding claim is that topic access earns enough buyer support to justify its effect on participation. @kahneman, audit that claim against the interviews and usage data.

3. Daniel Kahneman

   The managers threatening the renewal are vivid, so their answers carry too much weight. Start with the base rate: 19 of 28 employees expect to self-censor. A topic from a four-person team may identify someone after names are removed. Compare what participants did after each report type. Stated manager preference cannot settle the decision.

Deciding assumption

Topic access must earn enough buyer support to justify its effect on participation. Manager preference alone cannot settle the decision.

Use the challenge that matters

Observed behavior changed the call. Topic reports cut session completion from 72% to 49% for one extra manager vote.

1. Maya Chen

   We have that comparison. Teams with topic reports had 49% session completion. Matched teams with program reports held at 72%. Topic reports had support from 8 of 10 managers, compared with 7 of 10 for program reports. @socra, does one manager vote justify losing 23 points of completion?

2. @socra

   One more manager's support does not justify completion that is 23 points lower. The smallest report worth building shows progress toward agreed goals, attendance and kept commitments. It excludes conversation content and any view that identifies an employee.

3. Maya Chen

   The privacy model says department views still identify people because each department has only four to seven active participants. The release floor is ten contributors, and one person cannot supply more than 20% of observations. @socra, can we publish safely by combining departments?

4. @socra

   Only if every published cell survives a reconstruction attack. Publish only when at least ten active participants contribute and no person supplies more than one fifth of the observations. Exclude free text, topics, quotations and risk scores. Compare every new release with earlier reports and suppress any cell that could identify fewer than ten people.

5. Maya Chen

   Combining departments into four cohorts meets the ten-person floor. @socra, commit the counteroffer. What does Northstar get, what do we refuse, and when do we walk?

6. @socra

   Send a conditional renewal Friday. Northstar signs before we build. The 30-day pilot reports progress toward agreed goals, attendance and completed commitments for four combined cohorts. The report excludes topics, quotes, risk labels and individual views. The product has no raw export. The contract forbids employment use. The $180,000 renewal takes effect only if completion reaches 72%, at least 7 of 10 managers support the report, every cell passes the privacy test and the privacy reviewer cannot identify a person or infer a conversation topic. Walk if Northstar rejects the limits or refuses to sign first.

7. Maya Chen

   I will send the conditional renewal Friday and start the build after Northstar signs.

Conditional renewal committed

Counter before building. Report only progress, attendance and kept commitments for four privacy-safe cohorts. Exclude topics, quotes, risk labels, individual views and raw export. Walk if Northstar will not sign the limits first.

Decide what holds

The improved decision is ready for Maya to approve as Journey context. Proposed follow-up work remains a proposal until it is visibly created.

Approval required

The decision and follow-up proposal are ready for Maya to accept, revise, or reject.

• Northstar signs before implementation starts.
• Report progress, attendance and kept commitments for four combined cohorts.
• Exclude topics, quotes, risk labels, individual views and raw export.
• Release only after the 72%, 7-of-10, cell and inference checks pass.

Keep the signed-first rule, report scope, privacy exclusions and four release checks if they still represent the decision.

First action

Approve the decision context, then choose whether to create the report work.