How to interview for a role you do not understand
Learn how to assess a performance marketer or developer or accountant when you have never done the job yourself and cannot judge the craft.
Hiring outside your own craft
You need a performance marketer or an accountant or a backend developer. You have never done the job. Every candidate sounds plausible for the first twenty minutes and you have no reliable way to tell whether the confident one is also the competent one.
Reading about the craft before the interview does not fix this. Two weeks of articles will teach you the vocabulary candidates already use fluently. What fixes it is changing what the interview measures. You stop assessing knowledge you cannot verify and start assessing output you can.
What it costs and what it will not fix
This method does not fix a role you have not defined. If you cannot write down what the person must produce in their first ninety days then no interview technique will save the hire. It also does not replace a written scope or a salary band. Both of those come before the first call.
The direct spend is modest. A paid work sample on a real problem usually runs a few thousand rupees per shortlisted candidate. Paying an outside specialist to sit on one technical call costs their hourly rate. Both are invoiced professional services so GST at 18% applies on top of the fee. The real cost is attention. Designing one honest work sample takes an afternoon. Reviewing four of them properly takes another.
Who should use this and who should not hire at all
Some businesses should not run this interview because they should not be making this hire. If your sales numbers live in three spreadsheets and nobody agrees on last month's revenue then a full-time data analyst will spend six months cleaning up and then leave. Pay a contractor to fix the data first. The same holds for a first content hire at a business that has not settled what it sells and to whom. The hire becomes a very expensive way to discover you had a positioning problem.
The method suits an owner who has one recurring output clearly in mind and no internal person able to judge the craft. It suits replacement hiring for a seat that has existed before. It suits agency-to-in-house moves where you already know what good delivery looked like. It does not suit exploratory hiring where the brief is a job title and a hope.
How to run it
Start by writing the output rather than the skillset. Not "strong knowledge of paid media" but "owns the Meta and Google accounts and reports spend against enquiries every Monday". That single sentence does most of the filtering work.
Then build a work sample from a problem you actually have or recently had. Give every shortlisted candidate the same brief and the same four hours. Pay them for it. Ask for the reasoning alongside the deliverable because the reasoning is the part you can evaluate without craft knowledge. A candidate who explains why they rejected two other approaches is showing you judgement. A candidate who hands over output with no explanation is asking you to take it on faith.
Rank the samples before you look at the names. Then borrow a specialist for one call. A freelancer or a peer from another business will spot in fifteen minutes what you would miss in three interviews. Finally make reference calls specific. Ask what the person actually shipped and what they needed help with. Vague praise means the referee is being polite.
How to tell whether it worked
The metric that matters is whether you can rank the work samples with confidence and then watch that ranking hold up over six months. If your first choice is still your first choice at the half-year mark the process did its job. If the ranking inverts within weeks your work sample tested the wrong thing.
The metrics that flatter are application volume and how good the conversation felt. A long shortlist mostly reflects a generous salary band. A warm interview mostly reflects a candidate who interviews often. Neither predicts output.
The common mistake
The expensive error is skipping the paid work sample because it costs money and slows things down. A mis-hire in a role you cannot supervise is the most costly hiring mistake available to a small business. You pay salary through the notice period. You pay again in the months of your own attention spent discovering the gap. You pay a third time when the rest of the team loses confidence in how hiring is done here.
Measured against that a few thousand rupees per candidate and one borrowed specialist call is cheap. The owners who refuse the work sample are usually the ones who most need it.
A note on paid trial work
Paying a candidate for a work sample creates a commercial arrangement. Put the fee and the ownership of the output in writing before the work starts because an undocumented trial task leaves both sides unclear on who owns what. Treat the output as a test artefact and not as free production work. Using a rejected candidate's unpaid sample in live campaigns is where this practice earns its bad reputation. A qualified professional should confirm how this applies to your situation before you set it up as a routine.
This guide explains how something works. It is not legal, tax or financial advice, and rules change. Confirm anything regulatory with a qualified professional before you act on it.