Why completion rates matter
Take-home completion rate is a candidate-experience signal first, an evaluation signal second. In our data across 4,200 assignments, the completion rate ranges from twenty-eight per cent for the worst-designed take-homes to ninety-one per cent for the best. The difference is almost entirely about expectations, not difficulty.
If your completion rate is below sixty per cent, you are losing strong candidates to assignment fatigue. Start there before you tune the difficulty.
The two-hour cap
The single highest-leverage rule: give candidates a two-hour cap, and mean it. Tell them you grade what they finish in two hours, not what they polish in twelve. Completion rates jump dramatically when candidates trust the cap.
Companies that say 'spend as long as you like' implicitly select for candidates with twelve free hours, which biases the funnel. Our ATS ships the timer-based take-home as a first-class feature.
The eighteen rules
Rules in brief: include the rubric the candidate will be graded on, scope the problem to a single decision (not a system), provide all data the candidate needs (do not ask them to find data), make the deliverable format explicit (Markdown, Notion link, code repo, video), and never ask candidates to set up environments that are not part of the role.
The full eighteen are in our template library. They originated as internal hiring guidance and have travelled.
What changes for senior roles
For director-and-above roles, replace take-homes with a paid working session. The economics flip: senior candidates have less free time and less tolerance for the homework dynamic. Pay them for ninety minutes, structure the session around a real-but-anonymised company problem, and grade what you see in the room.
Structured interviews still apply; the take-home is just the wrong vehicle at that level.