How to assign a mock exam to a team, and what to do with the results

If you are putting a team through a certification, the most useful thing you can do early is sit them all down for the same mock exam in the same week. Not to rank them. To find out, cheaply, what the group does not know.

Why a cohort mock beats progress reports

Completion percentages tell you who has clicked through material. They do not tell you whether anybody could answer a question about it. A mock produces a score per exam domain for every person, and the overlap between those results is the single most actionable thing in a training programme: the domain where most of the team is weak is a training gap, and it is cheaper to fix once for everyone than six times individually.

Setting one up

In CertClue, an organization can create an exam for a cohort: pick the certification track, set the blueprint by domain, set a time limit and how many attempts are allowed, then publish it with an open and close window. Members in that cohort see it, sit it inside the window, and the board shows the results for that exam. The questions come from the same approved bank the rest of the product uses, drawn to the blueprint you set.

Two settings decide whether the result means anything. Weight the blueprint to the real exam's domain weighting rather than evenly, or the result will not resemble the paper anybody is going to sit. And allow one attempt for a diagnostic. Multiple attempts are for practice, not for measurement.

Reading the result

• Look at domain averages across the cohort first. One domain low everywhere is a syllabus gap, not six individual failures.
• Look at spread second. A wide spread on the same domain means one or two people already know it, which is a free internal teacher.
• Look at individuals last, and privately. A first diagnostic is meant to be poor. Treating it as a performance signal teaches people to game the next one.
• Ignore speed on a diagnostic. Finishing fast and finishing well are different things, and rewarding the first produces the wrong behaviour.

What comes after

1. Name the one or two domains the cohort is weakest on and say so out loud to the group. Shared weakness removes the embarrassment.
2. Give it a fortnight of focused study, with the material pointed at that domain rather than the whole track.
3. Ask whoever scored well on that domain to spend twenty minutes explaining it to the others. It is the cheapest and best-targeted training available.
4. Re-run a mock on the same blueprint and compare the domain averages, not the individual scores.
5. Book exams for the people whose second result clears the mark comfortably, rather than waiting for the whole group.

The thing to be careful with

A cohort mock is a training instrument. The moment it becomes an appraisal input, people optimise for the number, and the number stops telling you anything useful. If your organization needs an assessment of record, run that separately and say clearly which one is which.

It is also worth being explicit with the team about what the employer sees. On CertClue an organization sees results from the activities it sets and not a member's private study, their readiness score, or the practice they do on their own time. People study differently when they think everything is visible, and usually not better.