I gave away £490 of resources and started selling £49 exam packs

Here’s how the tutoring industry has changed in twenty years. One to one tutoring in person, then one to one tutoring online. Once people got comfortable with online tutoring, scaling it became possible. Some companies scaled by bringing in more tutors and holding the hourly rate. Others scaled one tutor by building online courses.

That’s what I did.

It was simple, really. Students liked learning from me. They didn’t want to learn from other tutors, so I couldn’t scale by hiring. That led me down the other route, and I set up GradePod. I recorded my best teaching material and sold it behind a paywall. For years I welcomed hundreds of students, and it was a hugely profitable and genuinely enjoyable business model.

Then in 2024 I started to notice sales dropping. People still wanted to learn from me, but they were less willing to spend £490 to get access to my resources.

I knew why. AI was starting to do a really good job of teaching students online, and they didn’t have to pay £490 to access it. They still wanted a human. But the job of personalised teaching was being handed to ChatGPT.

So I adapted. I added direct access to me inside the course, made it feel more like a community, and charged for the experience rather than the resources. It worked again, and it was a rather flattering reminder that people were interested in my personality and my teaching style as much as the material.

If I’m honest, I knew it wouldn’t last.

Habits change. Expectations change. ChatGPT has become a daily companion to students, and the instinct to come and ask me was always going to fade.

So in 2026 I changed the model completely. I made every previously gated resource, the ones valued at £490 a head, free on YouTube. Suddenly every student could learn from me and every one of them had the same access. It felt good to put access ahead of profit. But I still needed something to sell in a world that had stopped buying online courses.

Here is the one thing AI cannot do. It cannot make a student sit down and prepare for an exam the way a stack of past papers can. It is also the thing parents will happily pay for. If they think they can buy an exam pack, hand it to their child, and watch them work through twenty hours of it, that is money well spent.

So I built an exam pack that drives the practice of IB Physics rather than the learning of it, which is exactly what AI currently does badly. It is also completely passive for me, so I can price it low and sell more of it.

Each pack is £49. I have to sell ten of them to make what one person joining the course used to be worth.

I launched in March 2026, which was not ideal timing. The exams were in April, and after that I knew I had a fallow period until September, when schools go back and a new cohort of IB Physics students start thinking about the year ahead. Even so, I sold 151 packs in the first few months, which is not bad for the launch of a brand new product, on a brand new website, at the wrong season to launch. I have genuine confidence I can build these numbers up because YouTube is now my content engine, where Google SEO used to be.

Students watch the tutorials. They start to trust me. If they never buy the pack they still got something worth having, and if they do buy it they know they are buying from a human they already trust.

The channel has grown alongside it. At last check, 65,000 views, 6,200 hours of watch time and 637 new subscribers. It’s eligible to earn from advertising now, but that money is trivial and I don’t count it. The channel isn’t the revenue. It’s the reason anyone trusts me enough to buy the pack.

It feels like a better model all round.

I daresay online courses can still work, but they probably need the instructor to be all in: daily drop-ins, live lessons, maybe in-person events. I am honestly not that person, and my audience is worldwide. I cannot stay awake past 11pm running sessions for students in Australia.

What I do know is that passing physics exams is fast becoming a problem-solving exercise rather than a learning one. The best business models in this niche will help students become problem-solvers, hence the practice exam pack. The best business model for me is a passive one, hence the practice exam pack.