How compliance and completion are counted

Your reports carry two percentages that look like they measure the same thing. Compliance rate and Completion rate are counted differently, over different groups of people, and the gap between them is usually the most interesting number on the page. Nothing on screen explains the difference, which is what this article is for.

The two measures

Compliance rate counts only the people the course is mandatory for. Everybody else is left out of both halves of the sum, however many of them have taken it.

Completion rate counts everybody the course is assigned to, mandatory or not.

What counts as done

This is where they part company. Compliance counts somebody as done when they completed or passed. Completion counts them as done when they completed, passed or failed.

A failing score is a finished attempt, so it raises your completion rate and does nothing for your compliance rate. A course where many people fail will show a high completion rate beside a low compliance rate, and both figures are correct.

Only somebody’s latest attempt is counted. An earlier pass does not save them once they have retaken it and failed.

Attempts somebody abandoned

A learner who uses Start again leaves their previous attempt abandoned. An abandoned attempt has no result, so it counts towards neither measure.

If that abandoned attempt is their latest one, they stay in the group being measured while counting as not done, so both percentages fall until they finish the new attempt. That is the behaviour to expect after somebody restarts a course, and it corrects itself.

Completions you have superseded

When you publish a change and decide everybody has to take the course again, their earlier completions stop counting for the learner: they see the course as needing doing again.

Those completions still count in both percentages. The figures are not reduced by superseding, so a course you have just forced a retake on continues to report the compliance it had the day before.

This is worth knowing before quoting a figure in the week after a mandatory re-publish.

Taking an assignment away

Removing somebody’s assignment removes them from both measures entirely, rather than counting them as not done. Your compliance rate can therefore rise the moment you narrow an audience, without anybody completing anything.

People who reach a course through a learning path are counted against the path rather than against the course on its own.

How current the numbers are

These percentages are worked out in advance and stored, not calculated while you watch. Ordinary activity refreshes them as it happens, and a sweep every fifteen minutes catches anything missed.

That sweep only looks at accounts with activity in the last hour, so a quiet account is not recalculated continuously. A slower pass rechecks older figures and repairs any that have drifted.

In practice a figure you are looking at is current to within about fifteen minutes of the last thing that happened. It is not a live count, so a completion somebody has just finished may not be in it yet.

What this means for your setup

If you report on compliance, make the training mandatory. Optional assignments are invisible to that measure no matter how many people complete them.

When somebody challenges a figure, check the three things that move it without anybody completing anything: an assignment removed, an attempt abandoned, and a retake you asked for. Each has a legitimate explanation, and each looks like an error if you do not know it is there.