← Discovery hub
The network gets smarter for everyone — without anyone's data leaving.
Every practice contributes anonymised outcome rates — never records. The core finds what works, tests it against chance, and pushes only what survives back to practices like yours. A new site inherits all of it on day one.
Put your numbers in and the story below becomes yours.
Top-of-head is fine — nothing is saved or sent.
Illustrative reconstruction from the figures you typed. OakStone has no connection to your practice management system and holds none of your data. Every £ below is your arithmetic applied to a measured network effect — the +7.0pp rebooking result — not a measurement of your practice. The workings are shown so you can disagree with them.
Why it is worth more the longer you are in it — and as more join.
A finding can only be published once its stratum has five practices in it. So the network's value to you is not a flat fee for a tool: it is the share of strata that have cleared k≥5. Drag the network size and watch what opens up.
What each site contributes
Outcome rates, tagged to a stratum. Nothing else.
- A rate, not a record — e.g. rebook_rate 0.41 over n = 220 reminders.
- The setting that produced it — reminder offset, recall interval, sequence, wording variant.
- A stratum tag — region · size band · private-intensity band, so like is compared with like.
- A k≥5 gate — nothing is published or compared for a stratum with fewer than five sites. Below five it is suppressed, not shown.
Never leaves the practice: patient names, phone numbers, message content, clinical notes, or the identity of the practice itself.
How it comes back
A tested finding, as a versioned change you can audit.
- Only if it survives testing — meaningful effect size, a confidence interval that excludes zero, multiple-testing (FDR) correction, and minimum sample.
- As a config version — e.g. reminder_offset: 24h → 48h, v12 → v13, human-approved for anything material.
- Rolled back when it fails to replicate — monitored after release, reverted if the effect does not hold, and the reversal is reported to you too.
- Reported monthly — what changed, the measured effect, and every recovered £ ID-paired to two records in your own PMS.
A new practice inherits every validated finding on day one, then starts contributing. The advantage compounds from client one.
The validated finding that runs through the animation
Moving the reminder from 24 hours to 48 hours rebooked about 7 more patients in every 100.