It would look something like:
1. Use historical price data to determine the optimal price for each customer given what we know about them. 2. For a small number of customers (<10%), randomly adjust prices slightly to find additional data.
Once we have the data, we could find the price elasticity for customers (elasticity = % price increase / % change in purchases; so if elasticity is > 1, it makes sense to increase prices, if elasticity is < 1, you should decrease it), and use that to increase prices.
We did a few experiments via email, where we offered a different discount to different cohorts. For example, some people will get an offer for 20% discount, and others for 40%. We tried to extrapolate pricing elasticity from that, but it's still hard... We have a few ideas for other experiments, but didn't get round to it yet. It's one of the areas we need to be more careful.
I'd be happy to talk to you and try to understand what you have in mind. (I didn't quite). Hope it's ok to contact via the contact details on your profile? (Or feel free to reach out to me.).
We definitely want our service to be available to lots of students and to help them learn. But pricing it too low can not only leave money on the table, but actually make our product less valuable. Why? When you pay for something, you want to make the most of it. So our students have an even bigger incentive to keep learning. If it's too cheap, it loses its value, and in a way becomes less useful.
So we try to tier prices and offer a much lower price in developing countries for example. But we still want to find a "sweet-spot" for our main customer base, so people feel that they get value, but also that it's not too cheap so they stop caring about it.
Pricing experiments aren't always evil and the motivation isn't always to grab as much money as possible. Plus, don't forget that you can maximize profits by lowering prices in lots of cases. But you need to be able to test it to find out...