Skip to content

Product analytics — the prediction's measurement instrument

The brief's prediction says "grading time drops from 47 minutes to under 15." The analytics layer is what makes that checkable. Events are named after the person's action, not the system's operation:

text
// ✗ System events — what the code does
"button_clicked"      "page_loaded"       "form_submitted"

// ✓ Person events — what the person experiences
"grading-session-started"
"student-marked"
"grades-submitted-to-school"
"grading-session-completed"     // ← this is what measures the prediction

The time between grading-session-started and grading-session-completed is the prediction's measurement. The event schema traces directly to the brief. If the analytics layer doesn't fire these events, the prediction check in Volume V has nothing to measure against — and the cycle runs blind.

Next — Signals →

200apps · How We Work · NWIRE