Learning analytics was a genuine advance. It established that learning environments produce measurable phenomena, that those phenomena can be aggregated, and that institutions should look at them. That was not obvious twenty years ago.
But analytics inherited the shape of the systems it was built on. Learning management systems record transactions: submitted, viewed, logged in, completed. Analytics counted those transactions well. What it could not do was explain them.
01Three limits worth naming
- 01Proxy problems. Time-in-platform is not attention. Logins are not engagement. Submission is not understanding. Analytics frequently measures the shadow of the thing rather than the thing.
- 02Disconnection. The signals that would explain an outcome usually live in different systems than the outcome itself, and no one owns the join.
- 03Lateness. Descriptive reporting arrives after the window in which an intervention would have mattered.
02What comes next
Intelligence differs from analytics in what it is trying to produce. Analytics produces a measurement. Intelligence produces an interpretation with enough context and enough lead time to inform a decision.
The difference between analytics and intelligence is the difference between a record and an understanding.
That requires collecting different inputs — signals designed to carry meaning rather than transactions repurposed as evidence — and connecting them across the systems and timeframes where the actual story lives.
Artifact Research · Artifact Intelligence
Artifact Research publishes the working thinking behind the Learning Intelligence Platform, including the parts that are still open questions.