# Moat Verdict: Your Scheduling and Admin Platform ## Moat Scoring | Moat | Defensibility | Buildability | Compounding | Time to Signal | Total | |---|---|---|---|---|---| | Data | 5 | 5 | 4 | 3* | **17** | | Distribution | 2 | 4 | 3* | 4 | **13** | | Trust | 3 | 3* | 2 | 2 | **10** | **Data** — eighteen months of cancellation reasons, treatment codes and rebooking behaviour across 340 clinics cannot be bought, only accrued. **Distribution** — 70% referral inflow works today, but Facebook groups are rented ground any funded competitor can walk into next week. **Trust** — Cyber Essentials plus UK GDPR compliance is table stakes a competitor clears in a quarter, and rolling monthly contracts give you zero contractual lock-in. --- ## Primary Moat: Data **Verdict:** Your 12,000 bookings a month carry outcome data — why people cancelled, whether they rebooked — that no public UK physiotherapy dataset holds. Turn that from stored exhaust into a product feature that visibly recovers revenue, and a competitor needs eighteen months of live clinics before they can copy it. **Why it wins here:** - No-show patterns and rebooking behaviour at booking-level granularity exist nowhere else in UK physio; a competitor starting today has an empty table. - The asset is already collected and paid for — no new funding, no data purchase, only modelling work on what you hold. - Every one of the 340 clinics feeds the same behavioural model, so accuracy improves for all of them as you grow. - It attacks the weakest link in your business: rolling monthly contracts mean switching is easy, and a model that recovers lost appointment revenue makes leaving expensive in a way a contract cannot. **Trap moat: Trust** — It looks like the obvious play because clinics lose their entire appointment book if you go down and you hold patient health data. So you spend the quarter on ISO 27001, a formal SLA, penetration tests and an uptime page. At the end you own a certificate that a funded competitor buys in eight weeks, and you have not changed a single clinic's decision to stay. Security failure loses customers; security excellence does not win them. Keep Cyber Essentials current and spend the quarter elsewhere. --- ## 90-Day Build Plan | Window | Action | Expected outcome | KPI and target | |---|---|---|---| | Days 1-30 | Instrument every booking to close with a structured outcome code (attended, no-show, cancelled with reason, rebooked) | Clean training and measurement base | ≥95% of bookings carry an outcome code by day 30; source: bookings table completeness query | | Days 1-30 | Train a no-show risk score on the 18-month history; backtest on the most recent three months held out | Model that beats a naive baseline | Held-out AUC ≥0.70 and ≥20% lift over baseline no-show rate; source: internal eval notebook | | Days 31-60 | Ship an at-risk flag in the booking screen with one recommended action per flagged slot; run as A/B against a 25% holdout of clinics | Behaviour change, not just prediction | No-show rate ≥15% lower in treated clinics vs holdout; source: bookings table, split by cohort | | Days 31-60 | Ship a monthly Clinic Benchmark: this clinic's no-show and rebooking rate vs comparable UK physio clinics | Recurring reason to log in that only you can produce | ≥60% of the 340 clinics open the benchmark within 14 days; source: product analytics | | Days 61-90 | Surface recovered slots per clinic per month as a running counter in-app | Quantified switching cost visible to the buyer | Median ≥6 recovered slots per treated clinic per month; source: bookings table | | Days 61-90 | Publish an aggregate, anonymised UK Physio No-Show Index into the Facebook groups and the association listing | Data asset pulls distribution instead of replacing it | ≥25 signups attributed to the index; source: signup source field | --- ## Proof and Risk **Kill signal:** At day 90, treated clinics show under 5% relative no-show reduction against the holdout *and* under 30% of clinics opened a second benchmark report. That means the data describes behaviour but does not change it, and it is a reporting feature, not a moat. | Risk | Likelihood | Mitigation | |---|---|---| | Outcome codes stay incomplete because reception staff skip them | High | Default the code from calendar state; make manual entry a one-tap correction, never a form | | Clinics read the risk flag as blaming their patients | Medium | Frame every flag as a recommended action on the slot, never a patient score | | Association or Facebook group treats the published index as promotion and blocks it | Medium | Release aggregate findings with no product mention; let the source line do the work | **Assumptions made:** - Data / Time to Signal — no churn or engagement baseline was supplied, so signal speed is scored conservatively at one product cycle. - Distribution / Compounding — no referral rate per customer was supplied; assumed referrals are passive, not incentivised. - Trust / Buildability — no uptime record or incident history was supplied; assumed no differentiating reliability track record beyond the certification held.