Agreed, a "progressive" pricing based on salary would be a good way to address this, e.g. >50k -> 1%, >80k -> 3%, >100k -> 5%, >120k -> 7%.
The progressive^2 pricing model is interesting though. It's a little legally complex unfortunately so we have to look into it, but I also like the aspect of the idea where the less you make as a student, the more a lower percentage could help you live more comfortably.
Thanks to a focus on software and automation where we can afford it without hurting the experience, it's been good enough so far where we haven't noticed this affecting our decision making yet.
Another point: the cost per student does not increase linearly with time. Instead, advising becomes significantly more efficient after we've already taught the student our core insights. This means we're much more willing to spend additional time on students the more time we've already spent.
In contrast, houses can't "learn" how to sell themselves like a student can, so real estate agent costs are linear with time spent.