Turn test-rider feedback into evidence.

A rider logs a session in under a minute on their phone: which prototype, which conditions, a 1–5 rating per dimension, a voice note. That is the whole ask of the athlete, and it is deliberately small: the loop only works if logging a run is easier than sending a message about it.

Your R&D team sees every prototype compared against its production baseline, per dimension, with honest sample sizes. Rider quotes sit above and below the mean, so the numbers keep the sentence that produced them. A closed test plan is a verdict and a printable one-page readout for the shaper, not a pile of messages to re-read.

Multi-brand by design, brand-neutral by policy. Each brand sees only its own program. Nothing crosses between them, and that is a rule of the system rather than a promise about how we behave.

• Test plans. Define the experiment: candidate prototypes, baseline, riders, window, target conditions. Riders get pre-filled cards; you get coverage at a glance and a verdict when you close it.
• Comparison first. Prototype-versus-baseline deltas per dimension, trends and spreads, rider quotes attached, and a one-page readout you can print and hand over.
• Blind testing. Riders see "Front wing A". Only your team sees the code. Enforced server-side, not by convention.
• Custody and lineage. A QR passport on every prototype: scan to take custody or log a session, so you know who has what, since when, returned or damaged. V3 to V4 to V5 on one timeline with spec deltas and the quote that mattered.
• Rater calibration. Adjust for harsh and generous raters so a 3 from one rider means what a 3 from another does. Bring last season's spreadsheet and every row lands as a real, comparable session.

Where the hardware exposes logs, TestPilot can put what the rider felt beside what the board recorded on the same run. That is hard to add to a feedback tool afterwards, and it is the reason this product comes from us: the ride-data pipeline it needs already exists and already runs, moving sessions off Garmin watches and eFoil boards into comparable records.

The equipment side already exists too: a catalog of 225 products and which parts fit together, which is what lets a prototype be compared against a real production baseline rather than against a description of one.

What that buys you is narrower than a pitch deck would make it. Decisions rest on comparable data rather than on the loudest rider. A closed plan ends in a verdict instead of a thread to re-read. And the athlete program becomes searchable, rather than a chat history nobody can query.

Commercial

How a pilot works

A paid three-month pilot, including onboarding and a workshop to fit the rating taxonomy and questions to how your team already talks about a board. After that, a per-brand annual license. Athletes are unlimited, because charging per rider would push you to test with fewer of them, which is the opposite of the point.

TestPilot is in development and first pilots start in Q4 2026. If you run a test program and want to shape what it does, that conversation is more useful to us now than after it is finished.