Take-Home Tasks: When to Do Them and When to Decline

A well-designed exercise is the fairest assessment there is. It shows what you can actually do, rather than how well you interview, and it favours competence over confidence.

For a related Monitask resource on ways time trackers can be manipulated, see here.

A badly designed one is several days of unpaid work with a low chance of a job at the end.

The difference is measurable, and you can assess it before you start.

What a reasonable task looks like

Time-boxed and honest about it. Two to four hours, and the estimate roughly matches reality
Clearly relevant to the work you would actually do
Self-contained — hypothetical or anonymised data, not a live business problem
Explained — you are told what they are assessing and who will review it
Discussed afterwards, so you can explain your thinking
Positioned late in the process, after you have met people and know the role is real

What an unreasonable one looks like

No time estimate, or an estimate that is obviously understated
Days of work rather than hours
A real, current business problem. A marketing plan for next quarter, a fix for a live bug, a design for a product they are shipping
Requested very early, before any substantive conversation
No feedback, and no discussion afterwards
Deliverables they could use — finished copy, working code they own, a complete strategy document
"Show us how you'd approach our biggest challenge" with no boundaries

The clearest test: would the output be useful to them tomorrow if they did not hire you? If yes, that is work, and it should be paid.

Questions to ask before starting

Ask these by email. A reasonable employer answers all of them without friction.

How long should this take?
What specifically are you assessing?
Who reviews it, and will we discuss it afterwards?
What stage of the process is this — how many candidates, how many stages remain?
Is the data real or hypothetical?
Who owns what I produce?

Question four matters most. A task at final stage with two candidates is worth four hours. The same task sent to thirty applicants at screening stage is not.

Question six matters if the output has value. Reasonable employers confirm you retain ownership of work not used.

Deciding

Do it if: the task is time-boxed and relevant, you are far enough along that the role is real, and you would genuinely want the job.

Negotiate if: the task is reasonable but large. "I'd be glad to do this. Given the scope, would it work to do the first section and discuss the approach for the rest?" This is a normal proposal and is often accepted.

Ask about payment if: the scope is substantial and the output usable. "This looks like a full day's work — do you offer compensation for it?" Some employers do, particularly for senior roles. Asking is not presumptuous.

Decline if: it is days of unpaid work, or a live business problem, or requested before anyone has spoken to you properly.

How to decline without losing the role

Be brief, warm, and offer an alternative.

"Thanks for sending this over. Given the scope, I'm not able to take on a full unpaid project at this stage. I'd be glad to walk through my approach in a call, or share relevant work I've already done. Would either of those work?"

This is not confrontational, and it works more often than people expect. Employers who genuinely want you will usually accept a portfolio review or a working session instead. Employers who will not are telling you something.

Doing it well

Confirm the scope in writing before starting, including your time budget.

Stick to the time limit and say what you did. "I spent three hours, as suggested. With more time I would have added X, Y, Z" is a strong answer — it demonstrates judgement about scope, which is itself a competence.

Explain your reasoning. A short note covering your assumptions, what you prioritised, and what you would do differently with more information is often more valuable than the artefact.

State your assumptions explicitly. Most tasks are underspecified, deliberately.

Do not over-invest. Twelve hours on a four-hour task does not read as commitment. It reads as poor judgement about effort, and it disadvantages every candidate who followed the instructions.

Afterwards

Ask for feedback, whatever the outcome. You spent hours on it; a paragraph in return is a fair ask.

Follow up if there is silence. A task submitted into a void is a sign of a disorganised process, and worth knowing.

Keep what you made. Anonymise it and use it as portfolio material for the next application.

The wider picture

If a large share of the roles you apply for are asking for multi-day unpaid tasks, that is a pattern in your market worth noticing — and worth weighing when you decide where to spend your search time. Employers that assess well tend to manage well.

Looking for a role? Browse open positions → For broader reference, consult CareerOneStop.