Use the market data to structure a purchase, not to select a winner by price.
- Choose the format first. Decide whether the immediate need is coding foundations, sustained AI instruction, a bounded application or a research project.
- Normalise the quote. Separate live individual time, group time, asynchronous support and required extras.
- Approve a limited first commitment. Ask for a fit assessment and an early review point before a larger prepayment.
- Make learning evidence contractual. Request a short explanation of what the learner can now explain, change or debug, alongside the project demonstration.
- Make continuity visible. Agree how progress records, code, accounts and unused balances are handled if the mentor or provider changes.
For this desk, the highest-value next dataset is anonymised, comparable household quotations and renewal decisions. A proposed research panel should record quoted versus paid prices, delivered hours, age band, starting skills, project type, extras, refund terms and whether the family renewed. That would improve purchasing decisions more directly than another broad edtech forecast.
Sources and context
This page sets out the desk’s purchasing framework. How was this research conducted, and what are its limits? → Sources →