Why we don’t sell self-paced bundles
It would make the business easier in the short term and worse in the medium term. Here is the long version of that argument.
Once a quarter, a well-meaning advisor or investor asks us why Solutech does not sell self-paced bundles. The argument is usually some version of the following. You already have the curriculum. You already have the recordings. The marginal cost of a self-paced sale is approximately zero. You would broaden the funnel, reach engineers in time zones you cannot serve live, and make the unit economics of the business obviously better.
Every part of that argument is correct. We still do not sell self-paced bundles, and we are not going to. This is the long version of why.
What a cohort is
A cohort is not a delivery mechanism for content. We sometimes describe it as one, internally and externally, because that framing is easy to communicate. But the truth is that the content is the cheapest thing in a Solutech course. Recordings of the live sessions exist; instructor notes exist; reading lists exist. None of those things, individually or together, are what an alumnus pays for.
What they pay for is the part of the experience that is structurally impossible to bottle: forty engineers, in roughly the same place at roughly the same time, doing the same hard work and disagreeing with each other about it. The instructor adjudicates the disagreements. The peers force the work to ship. The schedule forces the schedule. None of those things survive the transition to self-paced.
We have tried. In 2024, we ran a quiet experiment with a small group of alumni and gave them a self-paced version of our flagship AI engineering course. The completion rate was roughly nine percent. The students who did complete it reported, with some chagrin, that what they had really wanted was the cohort.
The honest economics
People assume that the cohort model is a margin choice. It is not. Our gross margin on the cohort model is lower than it would be on self-paced; we pay instructors meaningfully more than the industry rate, and the cohort-size cap costs us obvious revenue. A self-paced bundle would be more profitable per sale.
What it would not be is more valuable per alumnus. Alumni who graduate from our cohorts go on to do specific things — promotions, founding their own teams, leading AI work at large organizations. That outcome is the long-term moat of our business. We trade short-term margin for long-term alumni outcomes, and we believe — based on five years of data we collect — that the trade is correct.
What we will do instead
We will keep doing what we have been doing. We will publish a small amount of writing for free — these essays, occasional course materials. We will run free office hours for engineers who cannot afford or cannot fit a cohort. We will share more of our curriculum publicly, on the theory that anyone who reads it carefully will see what is also true: that the material is necessary but not sufficient.
We will also keep saying no to investors who tell us we are leaving money on the table. They are correct that we are leaving money on the table. We are leaving the wrong customers on the table, on purpose. The right customers are the ones for whom the cohort is the value.
A small concession
There is one self-paced thing we are about to ship: a single, free, two-hour introductory module that any engineer can take in an evening. It is meant to make the question, “is a Solutech cohort right for me?” cheaper to answer. We expect a meaningful percentage of people who take it to conclude that the answer is no — and we will count that as a win.