- What reaches them
- A brief with a hard cap, submissions with the names stripped, and two review slots per submission on the team's calendar.
- What it means from there
- Reviewer minutes acquire a floor where they had none, and the loop is scheduled in advance rather than fitted around the week as it comes.
- What happens next
- Sets the scope so that three hours produces something gradable, and accepts the smaller sample as the evidence the hire is decided on.
Strict Take-Home Timebox & Blinded Evaluation Rubric
Enforce strict 3-hour scope limits on take-homes, compensate candidates for extended samples, and grade submissions blind.
Intervention I-006 Actor: hiring manager Scope: organizational Cost: medium Evidence: supported
On the canonical path
Assessed on the work, against a rubric that is shared — One timeboxed exercise that resembles the job. The candidate is given the criteria before starting, the timebox is enforced against the evaluator as well, and the same rubric scores everyone.
This entry is one of the things that holds that commitment up.
What it looks like when this intervention is in place
Composite excerpts, written to be typical rather than copied. No particular message, company or person is quoted here.
Title: Take-home — Senior Backend Engineer
Scope: one endpoint with persistence and tests. Hard timebox: 3 hours. Submit whatever you have at the three-hour mark.
Anything beyond three hours is paid at our contractor rate; tell us in advance if you want that option.
Review: submissions are anonymised before grading and scored against the published rubric by two reviewers, minimum 20 minutes each.
What to notice: Both halves of the asymmetry are addressed at once: the candidate hours are capped and the reviewer minutes have a floor. Capping only one side moves the problem rather than closing it.
What this targets
Expected effects
- Prevent reviewer fatigue and asymmetric code evaluation
- Respect candidate time investment and eliminate IP exploitation concerns
What to measure
take_home_completion_rateinter_rater_rubric_correlation
From each side
The same entry from inside each position that meets it: what reaches them, what it means from there, and what happens next given what they control.
- What reaches them
- A stated scope, a hard timebox, a rate for anything beyond it, and a published rubric the submission is graded against.
- What it means from there
- The hours are knowable before starting, and the grade is a scored artefact rather than an impression formed in a five-minute scan.
- What happens next
- Submits what exists at the three-hour mark instead of competing on hours, and weighs the task against the other processes in flight.
- What reaches them
- A request to pay a non-employee for an extended work sample, at a rate, on a contracting route.
- What it means from there
- Paying for a sample is a procurement and classification question before it is a hiring one; an unpaid task raises neither.
- What happens next
- Defines the rate and the route, or does not — and where it does not, only the timebox half of the intervention ships.
ATS vendorNot visible from here
From this position the entry is not visible. That is not a gap in the atlas — it is the finding: the decision is made where nobody in this position can see it.
Public policy and industry standardsNot visible from here
From this position the entry is not visible. That is not a gap in the atlas — it is the finding: the decision is made where nobody in this position can see it.
RecruiterNot visible from here
From this position the entry is not visible. That is not a gap in the atlas — it is the finding: the decision is made where nobody in this position can see it.
