Running a study contest for your team without wrecking the training

Week three is when team training dies. The kick-off energy is gone, the exam is still distant, and nothing bad happens to anyone who skips a week. A short contest is a decent answer to that specific problem, and a poor answer to almost every other one.

When a contest helps

It helps when the problem is momentum rather than understanding. If people know what to do and are simply not doing it, a visible board and an end date will move the behaviour. If people are stuck because a domain is genuinely hard, a leaderboard just tells them publicly that they are behind, which is the opposite of useful.

What to rank

Rank something that cannot be won by volume alone. A board that ranks on time spent rewards leaving a tab open. A board that ranks on questions answered rewards clicking through without reading the explanation. Rank on a game where the score depends on getting things right under a clock, keep the window short, and the incentive lines up with the training instead of fighting it.

In CertClue an organization can set a game contest for a cohort: pick the game, the track and the pace, publish it with an open and close window, and the board ranks the cohort on what they actually scored inside that window. Work done before the contest existed does not count toward it, which is what makes a new contest a genuine reset rather than a lap of honour for whoever started earliest.

The three ways it goes wrong

• It runs forever. Novelty is the active ingredient and it has a short half-life. A contest with no end date stops working in about ten days and keeps taking up space.
• It ranks the wrong thing. Anything that can be maximised by effort alone will be, and the people who do so learn nothing.
• It becomes a performance signal. The moment a manager mentions the board in a one to one, it stops being a game, and the quiet people who are already behind get further behind.

Who it is actually for

Contests reward the people who were already engaged. That is fine if you know it: the goal is to keep the middle of the group moving through a flat week, not to rescue the person who has not signed in since the kick-off. That person needs a conversation, not a leaderboard.

It is also worth saying out loud to the team what the board shows and what it does not. On CertClue a board covers the activity the organization set, and an employer does not see a member's private study or readiness score. People behave differently when they are unsure how much is visible.

A pattern that works

1. Run it in the third or fourth week, when the dip is predictable.
2. One game, one track, two weeks, published with a close date from the start.
3. Announce what it is for: keeping the habit alive, not deciding who is best.
4. Let it end. Do not extend it because it was going well; that is precisely when it has done its job.
5. Follow it with the next real milestone, usually a checkpoint mock, so the momentum lands somewhere useful.