Two big sites dominate. They're cheap monthly. They've been trusted for years. Most people prep in a two-month panic window. A new product from scratch — built by your dad, with AI — has no obvious right to win against that.

All of that is true. None of it is the whole picture.

• The entire market replaces itself every year. ~41,000 people sat the UCAT in 2025. Nearly all of them leave the market after one cycle — they get in, or move on. There is no locked-in customer base to fight for. Every January, the incumbents and everyone else start at zero with a brand-new cohort. "Dominance" in this market is one cycle deep.
• The incumbents can't copy our best moves. One-off pricing with no auto-renew would gut Medify's revenue model (their 3.1/5 Trustpilot is mostly people charged after their exam). Saying "the official free questions are better than any bank" would kill their own product. We can say and do both — they structurally cannot. That's not spin; that's the definition of an undefended position.
• Personalisation to last year's score exists nowhere. You checked — a re-sitter with a 1820 and a fresh Year 12 get identical content on every major platform. Re-sitter mode alone is a product category with zero competitors.
• Every current giant was once a newbie in an "owned" market. 6med was started by medical students. Medify launched into a market that already had established books and courses. New entrants win in this market roughly once a generation of applicants — because the customers are new every year (see fact 1).
• "Made with AI" is how everything gets made now. MedEntry ships an AI assistant. Medify is building AI features. The buyer's only question is "does it work?" — and if AI genuinely makes building this easy, it makes running it easy too. Worth thinking about who'd be well placed to run it.

Now price it from the buyer's side. The buyer is often a parent watching their kid chase medicine. What do they already pay? Tutors £50–100 an hour. A Medify summer £60–120 with an auto-renew trap. The test itself £70. An entire lost year if it goes wrong: unquantifiable. Against that, £79 once — for a system built around their kid's actual scores — is not a purchasing decision. It's a rounding error on the anxiety.

• Done, live now: the full training app, the free guide, the sales site, one-off pricing, refund policy. Built in a day; the incumbents' product cycles move in quarters.
• This cycle (now → Sep): free beta testers recruited from the applicant forums — re-sitters especially — in exchange for brutal critique. Their feedback hones it; the best of them become testimonials on results day (scores arrive same-day, receipts in hand).
• September: your result — whatever it is — gets published on the site. That's the founder story: real kid, real score report, before and after.
• January → next Sep: the 2027 testing cohort arrives with zero loyalty to anyone (fact 1), and we're there with proof, testimonials, and a price the big two can't match without breaking their own model.

• The daily cost is 60–105 minutes on weekdays, a 2-hour mock or analysis block at weekends, tapering before test day. That's not on top of your revision — it is the revision, structured. The only genuinely extra item is the error log: ~10 minutes a day, and it's the highest-value 10 minutes in the system.
• Keep Medify if you want it. The Engine runs alongside any question source by design — the drill timer doesn't care whose questions are on the other screen. Using both isn't disloyalty; it's actually the most useful test-bed data we could get: same student, both products, honest comparison from inside.
• It gives you hours back too: zero SJT grind (you're Band 1), VR held at maintenance (you're above the mean), and a stop signal so you never do the "one more mock" spiral.
• And the payback for test-bedding it: this thing is being built to sell from January, and the person who proved it should be paid from what it earns. The share is a conversation for you and Dad — but it's on the table in writing before your test, not as a vague promise after.