Take-Home Tasks: The Time Budget That Keeps Candidates in the Process

What the exercise is for

Guidance from 2023 is explicit: design assignments that measure between one and three specific skills, defined before the first candidate sees the brief. The temptation to load in extra requirements—to see how someone handles pressure, or to get a " fuller picture"—undermines the entire point. Each addition makes the signal noisier.

The same guidance requires a scoring rubric built before any submission arrives, applied identically to every candidate. Anonymous grading follows, stripping names from files to reduce the predictable biases that surface when reviewers know a candidate's background or source. These are process controls, not niceties. They exist because without them, comparisons across candidates collapse into impressionistic noise.

The hard constraint is singularity of purpose. If two finished submissions cannot be distinguished on the stated competency, the exercise is measuring something else—effort, free time, or tolerance for unpaid labor. That is not selection; that is filtering by circumstance.

The time budget is the test

Recruiting Daily caps assignments at two hours. This is not a suggestion. The two-hour limit forces compression: what can a candidate demonstrate in bounded time with bounded resources? The constraint shapes the task into something reproducible and fair.

The same source mandates that hiring teams complete the assignment themselves before release. Internal piloting catches the gap between stated and real time—between "should take an hour" and the three evenings it actually consumes. Without this step, companies advertise one commitment and extract another.

Completion windows matter too. Recruiting Daily recommends ten to fourteen days spanning at least two weekends. This accommodates candidates with full-time jobs, caregiving obligations, or non-linear schedules. A 48-hour deadline privileges the already-privileged.

The calendar window and the time budget are distinct levers. One governs when; the other governs how much. Conflating them—offering two weeks for a task that demands eight hours—is how companies lose candidates who read the requirement honestly and opt out.

Who drops out and why that matters

Interviewing.io tracked completion rates across engineering hiring funnels and found candidates finish take-home assignments approximately 62% of the time. The same source suggests that below 60% completion, the exercise likely needs revision. The company is not "screening for commitment"; it is hemorrhaging qualified applicants at a single stage.

The cost compounds downstream. Interviewing.io notes that losing more than 15% of candidates to attrition at any funnel stage signals process failure. A take-home that filters out 38% of starters before human contact is not a precision instrument. It is a bottleneck.

Market signaling reinforces the damage. Monster's 2026 jobseeker survey found 59% of respondents cite unpaid assignments or excessive take-home work as the factor most likely to deter them from applying. The exercise announces company values before any interview occurs.

Wolt Market Israel offers a concrete case. Calcalist Tech reported in July 2025 that 25% of candidates for management and junior positions at Wolt Market drop out specifically during the assignment stage, a figure given by the company's Israel chief executive, Shmuel Navon. The figure is internal and specific: length and ambiguity drive abandonment, not lack of interest or skill.

How to make scoring consistent

Rubrics must precede grading. Recruiting Daily's instruction is procedural: define criteria before viewing submissions. The alternative—improvising standards after seeing the range of responses—introduces variance that masquerades as discernment.

Anonymous review follows. Names and identifying details stripped from submissions prevent the predictable associations that distort evaluation: the prestige of prior employer, the familiarity of educational pedigree, the demographic cues embedded in presentation choices.

What remains unsourced in the record is how to handle borderline submissions. Some hiring guidance recommends a second independent reviewer for borderline cases, but this is not verified in the sourced materials. Hiring teams without established inter-rater reliability checks are flying blind on consistency.

What candidates need to know up front

CreateCV's candidate guidance, published August 2026, documents what applicants are advised to demand: an approximate hour count, not merely a calendar deadline; clarity on when the timer starts and stops; confirmation of deliverable format, length, presentation requirements, intended audience, and the specific factors reviewers will weight.

These are not negotiating points. They are the minimum viable specification for a task that respects candidate time. A company that cannot answer them has not designed the exercise; it has outsourced labor assessment to hope and inference.

The asymmetry is structural. Candidates bear the full cost of ambiguity—extra hours, wrong assumptions, rejected submissions that missed unstated criteria. Employers bear none. Correcting this requires explicitness: time boundaries, scoring transparency, and a scope that fits the two-hour ceiling.

If the task cannot be defined within these constraints, the honest alternative is to abandon it. The brief's final section on replacements—pairing sessions, work walkthroughs—lacks direct sourcing. Until primary research compares these methods directly, hiring teams must judge from first principles: whether synchronous evaluation or portfolio review answers their specific question with less friction and equivalent validity.

The obligation runs one direction. When a company asks for hours of unpaid labor, it owes candidates precision: what, how long, how judged, and to what end. Anything less is not a hiring process. It is a tax on applicants' time, collected without consent and rarely receipted.