6The measures
We decide what will be measured before the first course is built. The data can then diagnose a course while it is live, instead of explaining a failure afterwards.
You bringThe results you report on today, and who needs to see them.
You leave withAn evaluation plan: what is measured at each level, where the data comes from, who reads it.
In the StudioEvery exercise is tagged with the skill it tests. Answers flow to the CRS and the Data Hub, and the Auditor turns them into what to fix next.
What AI does hereAI reads the answers at a scale no team can, and points to the courses and questions that need work.
- Data-driven learning design diagnose, don’t autopsy. Every click and drop-off is digital body language.
- Kirkpatrick’s four levels reaction, learning, behaviour, results.
- Learning analytics source the data, arrange it, monitor, find the insight, act.
Worked example · a general language curriculumEach answer is recorded as recognising, producing or recalling. The first-attempt score and the review at the end show what was retained.