Careers

The Paid Work Trial

Some time ago I gave a candidate a take-home assignment that I had estimated at two hours. He sent it back with a short note saying it had taken him most of a weekend. He wasn't complaining. He thought I should know, in case I was calibrating future assignments against his. I had no good answer for him. I had designed the exercise, priced his time at zero, and then been surprised that zero did not buy two hours.

The unpaid take-home is one of those practices that spread because each step looked reasonable. An interview is a poor predictor of how someone works, so we ask for a sample. A sample is more informative if it is realistic, so we make it bigger. Nobody stops to ask who is absorbing the cost of that growth. The answer is the candidate, and not all candidates equally. The person with a free weekend does better than the person with a full-time job and two kids, and we mistake that gap for a gap in ability. The exercise ends up measuring spare time as much as skill.

What a trial measures that an interview cannot

An interview measures how well someone talks about work. That is related to doing the work, but not the same thing, and I have been fooled in both directions. I have hired articulate people who stalled when the brief was messy, and passed on quiet people who would have been excellent. A paid work trial puts a person inside a small, real problem for a short stretch, and you learn things no conversation surfaces. You see what they do when the brief is ambiguous. You see whether they ask questions early or guess and hope. You see how they take a correction at four in the afternoon.

In the best trial I have run, the most useful signal was not the deliverable. It was a question the person asked within the first hour, one that exposed an assumption in our brief that we hadn't noticed. The finished work was good, but the question told me how they would think on a Tuesday six months later, when nobody was watching.

How I would design one now

The rules are simple. The trial is paid, at a fair rate, whether or not we hire the person. It is bounded: a fixed window and a defined scope, written down in advance. The problem is real enough to be meaningful but contained enough that nothing depends on the result. And a specific person on our side is available to answer questions, because how a candidate works with us is half of what we are trying to learn.

I got that last part wrong the first time. I wrote a vague brief, stepped away, and then marked the candidate down for seeming confused. The confusion was mine. I had built a test of how well someone could read my mind and then treated the results as information about them. Now I read the brief aloud to a colleague before anyone sees it, and if either of us hesitates, it gets rewritten.

The objections I take seriously

The obvious objection is that people who are currently employed cannot give up days for a trial, and that is true. A paid trial helps, but it does not create time. For those candidates I have offered a smaller scope spread over evenings, or a single paid half-day working session. I do not know whether that fully solves the problem. It narrows the gap without closing it, and I would rather say so than pretend I have a clean answer. For very senior roles a trial may not be the right instrument at all, and a deep conversation about work they have already done may serve better.

The other objection is cost. It is real, and it is small next to the cost of a wrong hire, which I have paid and which was larger than I admitted at the time. But I would not defend the practice on arithmetic alone. Paying someone for their work is simply the correct baseline, and it happens to improve the quality of what we learn, because people do their real work when they are not gambling their weekend on a maybe.

There is a side effect I did not anticipate. A trial lets the candidate evaluate us. They see how fast we answer, whether our feedback is specific or vague, and whether we respect the boundary we set. Several of the strongest people I have worked with said afterward that the trial was what made them want the job, because it showed them what working here would feel like. That is the same belief behind Provieo: hiring should rest on demonstrated work, and the path to it should not require anyone to donate their time. My rule now is plain. If I would not pay for the work, I should not ask for it.