Feedback
The loop closes here. Measurement without action is theatre.
Kill criteria and the scorecard are defined in the business plan. This page tracks five KPIs that drive decisions — not activities that feel productive.
Scorecard Summary
| Dimension | Score | Evidence | Conviction |
|---|---|---|---|
| Pain | 4/5 | Isolation epidemic documented, AI accelerating disconnection | HIGH |
| Demand | 2/5 | Zero paying customers, touch norm varies by context | LOW |
| Edge | 3/5 | Measurable trust delta differentiates from entertainment | MEDIUM |
| Trend | 4/5 | AI compressing cognitive work increases value of embodied trust | HIGH |
| Conversion | 1/5 | Zero conversion data. Pre-pilot. | NONE |
| Composite | 96 | Below 100 threshold — needs demand evidence to proceed at scale | — |
Quadrant: INVEST — High priority + low preparedness. Close the preparedness gap (pilot evidence, facilitator recruitment) before scaling.
Five KPIs
MRR is not on this list. MRR is an outcome of these five working together. Track MRR for accounting. Track these five for decisions.
| # | KPI | What it measures | Why this one | Target |
|---|---|---|---|---|
| 1 | Trust delta per session | Pre/post validated trust scale score | North star. If the protocol doesn't move the dial, nothing else matters. Zero delta after 5 cohorts = protocol doesn't work. | >0.5 standard deviation improvement per session |
| 2 | Repeat attendance rate | % of participants returning for session 2+ | Proves the experience created belonging, not just novelty. Commitment and consistency — people return to what resonates. | >50% by Month 3 |
| 3 | Facilitator hours per session | Actual prep + delivery + measurement hours | The kill metric. At 60% gross margin, 2x hours = 20% margin = below meaningful return. | <8 hrs/session total |
| 4 | Touch opt-out rate | % of participants who opt out of physical contact elements | Safety signal. High opt-out means the context needs a no-touch protocol variant, not abandonment. | <30% opting out of all touch elements |
| 5 | Facilitator retention | % of certified facilitators still running sessions after 3 months | The network effect metric. If facilitators don't stay, the distribution model doesn't work. | >70% still active at 3 months |
Why not these
| Metric | Why it's excluded |
|---|---|
| Session attendance numbers | Vanity. 50 attendees and zero trust delta means the format doesn't work. KPI #1 and #2 tell you the truth. |
| Social media mentions | Activity, not outcome. Trust isn't built or measured in social feeds. |
| MRR | Outcome, not driver. If the five KPIs are green, MRR follows. If MRR is green but KPIs are red, it won't last. |
| Facilitator certifications issued | Output, not outcome. A certified facilitator who doesn't run sessions contributes nothing. KPI #5 tracks what matters. |
Off Course — Decision-Action Process
Bad results don't mean do more. They mean ask better questions. Every red signal is data about which assumption was wrong.
| KPI | Red signal | What it indicates | Decision-action |
|---|---|---|---|
| Trust delta = 0 | 5 cohorts, trust scores unchanged | Protocol doesn't create the mechanism | Don't run more sessions. Redesign the opening ritual first — it sets the social contract. Run 3 interviews with participants: "What would have needed to happen for you to feel more connected?" If language differs from your protocol design, rewrite. If the mechanism isn't there, thesis is wrong. Trigger kill criteria #1. |
| Repeat < 30% | 3 cohorts, majority don't return | Session was interesting but not compelling | The experience is entertaining, not infrastructure. Check: was it the ritual design or the facilitation quality? Run a comparison — same protocol, different facilitator — to isolate the variable. If design is the issue, redesign. If facilitation, double down on QA. |
| Facilitator hours > 12 | First 5 sessions averaging 150%+ of estimate | Protocol not templated enough, too much custom prep | Don't take more facilitators. Stop. Document which hours are repeatable vs custom. If >50% custom, the "protocol" is improvisation — not infrastructure. Templatise or reprice. |
| Opt-out rate > 60% | Majority avoiding touch elements | Context doesn't support the touch premise | Don't force it. Run a no-touch variant immediately. If no-touch variant also produces positive trust delta, the touch elements are not the mechanism — the ritual structure is. That's a different (possibly better) product. |
| Facilitator retention < 50% | Facilitators trying once then stopping | Platform not supporting facilitators, or protocol too hard to run | Interview every facilitator who stopped. Map the friction points. If the protocol is too complex for sustained facilitation, simplify it. If the support is missing, build it. A protocol nobody wants to run is not infrastructure — it's homework. |
On Course — Amplification Process
Don't celebrate. Capture. The window between "this is working" and "too busy to document why" is short.
| KPI | Green signal | What it indicates | Decision-action |
|---|---|---|---|
| Trust delta > 1.0 | Consistent strong trust improvement | Protocol mechanism confirmed | Document exactly which elements drove the change. Publish the before/after data (with participant consent). This is the primary evidence for every sales conversation. |
| Repeat > 60% | Strong return rate | Experience creates genuine belonging | Don't change the ritual. Identify what participants say when they describe why they returned — that language is your positioning. Start a waitlist. Scarcity signals quality. |
| Hours < 6 | Templates working, protocol efficient | Capacity to scale facilitators | Don't fill capacity. Use spare hours to build facilitator training materials. Record one full session as training video. This is how you certify the first facilitator cohort. |
| Opt-out < 15% | High consent alignment | Context is right, protocol design is working | Document the session setup that produced this result — location, group composition, facilitator approach. This is the recipe for replication. |
| Facilitator retention > 80% | Facilitators staying and running sessions | Network effect thesis showing early signs | Stop recruiting broadly. Go deep with retained facilitators — what do they need to run more sessions? Give it to them. The network builds through depth, not breadth. |
Measurement Log
| Date | KPI | Planned | Actual | Gap | Action |
|---|---|---|---|---|---|
| 2026-03-18 | Phase completion | Protocol seed set | In progress | — | Define first 5 protocols |
| — | Trust delta | >0.5 SD | — | — | — |
| — | Repeat attendance | >50% | — | — | — |
| — | Facilitator hours | <8 hrs/session | — | — | — |
| — | Opt-out rate | <30% | — | — | — |
| — | Facilitator retention | >70% at 3mo | — | — | — |
Loop-Back Trigger
| Signal | Loop back to | Action |
|---|---|---|
| Touch elements rejected in target context | VALIDATE | Run a no-touch protocol variant. Re-score Edge dimension — the mechanism may be ritual, not touch. |
| Facilitators won't certify | MODEL | Shift from facilitator network model toward direct service delivery. Revisit Stage 2 timeline. |
| Trust delta proven, nobody pays | STRATEGY | Positioning problem. What kind of buyer values proof over experience? May need to move upmarket to corporate L&D. |
| Delivery hours 2x estimate | FINANCE | Reprice or reduce scope per tier. Facilitator training must reduce marginal hours. |
| Opt-out rate high but trust delta positive | DEFINE | The mechanism is not touch — it's ritual structure and consent process. Rename and reposition. |
| 90-day proof plan fails at any gate | KILL | Follow kill process in business plan. Extract learnings into protocol template library. |
Questions
Which of the five KPIs will produce its first real number — and what decision does that number force?
- If trust delta is strong but repeat attendance is weak, does that validate the mechanism or invalidate the product?
- At what point does a green facilitator retention rate signal that Stage 2 certification investment is safe — one quarter or two?
- If touch opt-out is high but trust still improves, have we discovered a different (and more scalable) product?