All documentation

Class engagement metrics

Which students are still with you, and which are quietly on their way out.

Tutors & Institutions8 min read
Who this is for
Tutors and institutions running batches, who need to know who is disengaging while there is still time to do something about it. Open it at Dashboard → Engagement.

Overview

Students rarely announce that they are dropping out. They attend a little less, stop submitting, go quiet in chat, and then one day they are gone — and the first anyone notices is when the renewal does not come.

Engagement metrics make that visible while it is still reversible. The page reports, per batch and per student, whether people are turning up, handing work in, and taking part.

Batch-level metrics

Pick a batch and you get its headline numbers:

MetricWhat it tells you
Total studentsHow many are enrolled in the batch.
Total classesHow many live classes have been held.
Average attendance rateAcross the batch — the single best early warning.
Average assignment completionWhat proportion of set work is actually coming back.
Total assignmentsHow much you have set, which is the denominator for the above.

Batch-level numbers answer “is this class healthy?” A single weak student in a strong batch is a conversation; a whole batch at 45% attendance is a scheduling or pacing problem you own.

Student-level metrics

Underneath, each student is listed with their own numbers:

  • Attendance — classes attended against classes held.
  • Assignment completion — submitted against set.
  • Chat participation — whether they say anything in class.
  • An overall engagement score combining those.

Sorting by this is how you find the three students to contact this week rather than reading forty rows.

The engagement score

One number per student, combining attendance, assignment completion and participation, so a batch can be ranked at a glance. It is a triage aid, not a grade: a quiet student who attends everything and submits everything is not disengaged, they are quiet.

Reading the colours

BandRangeWhat it means
Green75% and aboveEngaged. No action needed.
Amber50–74%Slipping. This is the band where a message still works.
RedBelow 50%Disengaged. Needs a direct conversation, and possibly a parent one.
Amber is the band that matters
By the time a student is red they have usually decided. Amber is where intervention is cheap and still works, and it is also the band that is easiest to scroll past.

At-risk students

Alongside the metrics, students are flagged as at risk when either is true:

  • Attendance has fallen below 60%, or
  • They have two or more overdue, unsubmitted assignments.

Two separate triggers, because the two failure modes look different. A student who attends but never submits is struggling with the work; a student who submits but stops attending has a timetable or motivation problem. The right conversation is not the same one.

How each number is measured

MetricSource
AttendancePresence in your live classes for that batch.
Assignment completionSubmissions against assignments set to that batch — see assignments & grading.
Chat participationMessages sent in class chat, used as a proxy for taking part.
Chat is a proxy, not a truth
Class chat activity stands in for participation because it is the signal that exists. A student can be fully engaged and silent — treat low participation as a prompt to look, not as a finding on its own.

What to actually do about it

  1. Sort by engagement and look at the bottom five. That is the whole weekly ritual.
  2. Message the amber ones directly. A specific message — “you missed the last two, is the timing hard?” — outperforms a batch announcement by a distance.
  3. Check whether it is you. If a whole batch dips in the same week, look at what you set or scheduled that week before looking at the students.
  4. Escalate red to a parent where there is one linked — see parent-teacher messaging.
  5. Cross-check with results. High engagement with poor scores is a teaching problem, not an attendance one. Student insights has that side.

Limitations

  • Only live classes count toward attendance. A batch taught mostly through recorded content will show low attendance without anyone having disengaged.
  • Participation measures chat, which under-reports students who engage by listening.
  • A batch with no assignments set has no completion signal at all.
  • Numbers are only as good as the batch setup — students in the wrong batch produce misleading rows.

Common questions

Do students see their engagement score?
No. This is a teaching view. Students see their own progress through their own dashboard.

Why is attendance zero when the class ran?
Attendance is recorded from the live class itself. A session held outside the platform cannot be counted.

Can I see engagement across all my batches at once?
Metrics are reported per batch, which is the unit you actually teach and intervene in. Move between batches with the selector.

Does this affect a student's grade?
No. It is a teaching signal only, and it feeds nothing the student is assessed on.


Header Logo