Almost every household we deal with agrees in principle that a trial is sensible, and a surprising number skip it anyway — because the need is urgent, because the candidate seemed obviously fine, or because nobody was sure how to run one. The result is a hire made on an interview, and an interview tests how well somebody interviews.
What an interview can and cannot establish
An interview establishes whether somebody presents well, answers questions coherently, and claims relevant experience. Those are not worthless signals. They are also not the job.
What an interview cannot establish is any of the things that actually determine whether the arrangement works: whether he can cook food your household will eat, whether he sequences a set of tasks sensibly, whether he cleans up after himself, whether he asks when he does not know, how he handles an instruction that changes mid-task, and what your kitchen or your car looks like at the end of a day in his hands.
Every one of those is visible in a single paid day and invisible across three interviews.
What a paid trial costs and what it saves
A trial costs a day's wage, sometimes two. That is the entire cost.
A failed permanent hire costs a month or two of wages, the disruption to the household while it unravels, the awkwardness of ending it, and the time spent hiring again. For a household relying on staff for a school run or an elderly parent, it costs more than that.
The ratio is not close, and it is the reason we insist on trials for every placement rather than offering them as an option.
Paid, not unpaid — and why this matters
An unpaid trial systematically selects for the wrong candidates.
A capable person currently employed elsewhere cannot take an unpaid day. He has a job. The candidates who accept unpaid trials are disproportionately those with nothing else on, which is not a pool you want to restrict yourself to.
It is also simply work. A man who cooks a meal for your family, or cleans your house for six hours, has worked, and paying for it is the baseline of being a decent employer. Households that begin an employment relationship by taking a day's labour for free have set a tone that will shape everything after it.
Pay the going daily rate. If he is not suitable, you have spent one day's wage learning it. If he is, you have started well.
How to design a trial that actually tells you something
The principle is that the trial should be an ordinary day, not a demonstration.
For a cook: ask for a normal weekday meal the household actually eats, for the real number of people. Not something impressive. Watch the sequence, the quantities against the number at the table, the oil and ingredient use against what you budget, and the state of the kitchen at the end. Whether he asks about preferences is the single most predictive thing you will observe.
For a helper: give him a written list of the real morning's tasks in the order they would happen, and watch sequencing — whether he finishes one thing before starting another, whether he cleans up the tools of one task, and whether he reports what he could not complete.
For a driver: a full day on your actual routes in your actual vehicle, including a school gate or a busy market if those are part of the job, and at least one route he does not know so you can see how he handles not knowing.
For a guard: a written scenario exercise — a delivery, an unknown visitor, a known visitor when nobody is home, an incident — and look at whether he can produce an accurate written report.
For a caretaker: walk him round the property and ask what he notices. A real caretaker points at things you did not mention.
Our paid practical trial guidance sets out the structure for each role in more detail.
The mid-task change, which is the best single test
Whatever the role, introduce one change part-way through the trial. Not to trip him up — to see what he does.
Ask for something additional while he is in the middle of a task. Change a timing. Add a person to the meal. Redirect a journey while safely stopped.
What you are watching for is whether he asks a clarifying question, whether he tells you what will be delayed as a result, or whether he simply abandons the first thing and does the new one silently. Households run on constant small changes, and how a person handles one is the closest thing to a prediction of the next three years that a single day can give you.
Two trials, a week apart, for roles that matter
For a cook, and for any role where consistency is the point, one good day proves less than people think. Anybody can produce one good meal.
Two paid days a week apart, cooking different things, tells you whether the standard is repeatable. It also tells you something about reliability, because the second appointment has to be kept.
For high-exposure roles — live-in positions, care of an elderly family member, a driver carrying children alone — the extra day is worth considerably more than it costs.
The probation period after the trial
A trial is not the same as probation, and both are useful.
After a successful trial, the first month should be an explicit probation with a review at the end of it, and both sides should know that. Pay weekly during that month if you like; it gives either party an easy exit without a month's money hanging over it.
Then hold the thirty-day review properly: what is working, what is not being done at all, what needs to come off the list because the day is genuinely full, and whether the salary matches the job as it has turned out to be. Our thirty-day review guidance covers the short version.
Households that hold that conversation once keep staff for years. Households that skip it are the ones surprised by a resignation in month five.
What to tell the candidate
Be straightforward, because vagueness here is what gives trials a bad reputation among workers.
Tell him it is a paid trial, what the rate for the day is, what he will be asked to do, how long it will take, and when you will tell him the outcome. Then tell him the outcome when you said you would, including if the answer is no.
Candidates in Islamabad have generally been through trials that were unpaid, open-ended, or never followed up. A household that runs one properly is immediately a more attractive employer, and good candidates notice.
Islamabad conditions worth knowing
The candidate pool here is deep enough that you can afford to trial two or three people for a role rather than one, and for an important position that is worth doing. It costs three days' wages and it converts hiring from a guess into a comparison.
It also means good candidates have alternatives. A trial that drags on for a week unpaid, or a household that takes ten days to decide, will lose the better candidates to employers who move faster. Decide within a day or two.
For live-in roles, run the trial before anybody moves in. Reversing that order makes an unsuccessful trial into an eviction, which is unfair to him and unpleasant for you.
Trialling more than one candidate
For an important role, running two or three trials rather than one converts hiring from an accept-or-decline into a genuine comparison, and the difference in outcome is considerable.
Use the same day's tasks for each candidate so you are comparing like with like. Take five minutes afterwards to write down what you observed while it is fresh — sequencing, cleanliness, questions asked, the mid-task change — because after the third candidate the days blur together and impressions replace observations.
Then compare on the things that predict the next three years rather than on likeability. Did he ask about the household's preferences? Did he report what he could not finish? Did he leave the work area in order? Likeable candidates who did none of those are the ones households regret.
Tell each candidate the timeline and honour it. In a city where good staff have alternatives, a household that takes ten days to decide loses the person it wanted to somebody who moved faster.
When to skip the trial
There are two situations where going straight to a start is defensible.
The first is genuine emergency cover — a driver needed tomorrow because the school run cannot lapse, a cook for a fortnight while yours is away. Here the honest approach is a short, explicitly temporary engagement with weekly pay, treated as an extended trial rather than as a permanent hire, with the permanent decision made later.
The second is a candidate coming from a household you know personally and whose work you have actually seen. Even then, one day in your own house on your own routines is worth having, because the job is different from the one he was doing.
What is never defensible is skipping the trial for a live-in role, for care of an elderly family member, or for a driver who will carry children alone. Those are precisely the positions where a bad hire is not recoverable, and a day's wage is a small price for knowing.
Frequently Asked Questions
What is a fair rate for a trial day?
Roughly a day's share of the monthly salary you are offering for the role. Rounding up is a good habit and costs almost nothing.
Can I trial somebody while my existing staff member is still working?
Yes, and it is often sensible for continuity. Be straightforward with both about what is happening — running a trial in secret while somebody is working out his notice creates a bad atmosphere that the new person inherits.
What if the trial goes well but I am still unsure?
Run the second day. Uncertainty after one day is usually about something specific — pace, communication, a particular skill — so design the second day around exactly that question.
Should I write the scope before or after the trial?
Before. The trial should be a day of the job as you have defined it, which is only possible if you have defined it. Our hiring checklist covers what that page should contain.
