Why Interview Questions Don't Predict Admin Performance

The admin interview measures conversation, not competence. Here is why the standard questions fail, what they actually reveal, and the ten-minute step that fixes the gap.

"Tell me about a time you had to manage competing priorities." "How do you make sure nothing slips through the cracks?" "Describe your approach to attention to detail."

Every admin interview contains versions of those questions, and every candidate has an answer ready. The answers are usually good. They are also almost entirely uninformative, because the question asks the candidate to describe a skill, and describing a skill and having it are different things. This page explains why the admin interview fails as a predictor, what it is actually good for, and what to add so the hiring decision rests on evidence instead of impression.

What an interview measures

An interview is a structured conversation. It is good at measuring the things a conversation reveals: whether the candidate is articulate, whether they are pleasant to spend forty minutes with, whether their story hangs together, whether they have thought about the role. Those are real qualities and some of them matter.

But the core of an administrative job is not conversation. It is receiving instructions and executing them exactly, holding details, catching errors, applying rules. An interview never asks the candidate to do any of those things. It asks them to talk about doing them. So the person who is good at talking about attention to detail scores well, whether or not they have any.

The research on this is old and consistent. Unstructured interviews are among the weakest predictors of job performance of any common selection method. Structured interviews do better, and work samples, where the candidate actually performs a piece of the job, do better still. For admin roles in particular, where the job is a sequence of small, concrete, checkable actions, a short work sample is the closest thing to watching the person work before you hire them.

The four standard questions and what they really tell you

"How do you stay organised?" You learn whether the candidate has a vocabulary of organisation: lists, calendars, colour codes, inbox zero. You do not learn whether, given three day-and-task pairs on a call, they can tell you an hour later which task was on Wednesday.

"Tell me about a mistake you made and what you learned." You learn whether the candidate has prepared a safe mistake. Everyone has one. It is usually from a previous job, involved a lesson about communication, and had no lasting consequences.

"How do you handle multiple instructions at once?" You learn that they prioritise, write things down and confirm back. You do not learn whether they can carry out a four-step instruction with step two moved to the end, because you did not ask them to.

"Why do you want this role?" You learn something about motivation, which is genuinely useful, and nothing about skill.

None of these questions are bad. They are just measuring a different thing from the thing that decides whether the hire works.

Why the gap matters more for remote roles

In an office, the interview's weaknesses are partly covered by what happens after it. You see the new hire's first week. You notice whether they write things down, whether they ask when unsure, whether their first outputs are clean. A wrong hire is often visible in days.

Remotely, especially across time zones, those signals arrive late or not at all. You see finished work, not the process. A missed step surfaces when the client calls. By the time the pattern is undeniable, you have paid for two or three months. So the interview, already weak, carries a larger share of the decision in exactly the setting where its weaknesses hurt most.

Offshore hiring sharpens this further. The candidate pool is large and competitive, which pushes every applicant to present polish: a professional profile, a confident interview, fluent answers to all four standard questions. That polish is a rational response to the market and no reflection on anyone's character. It does mean that the interview, as a discriminator between candidates, is close to useless, because everyone clears it.

The fix: add a work sample, keep the interview

The answer is not to abandon the interview. It is to stop asking it to do a job it cannot do.

Keep the interview for what it measures: motivation, communication, whether you would want to work with this person. Then add a short work sample for what it does not: can this person, today, do the small concrete things the role requires?

A good admin work sample has a particular shape. It is short, around ten minutes, so the candidate's goodwill and attention are not the thing being measured. It is made of tasks that look like the job: a multi-step instruction with a twist, a short schedule to remember after it disappears, a sentence with a wrong date to correct, two numbers to add, a simple rule to apply. It is taken on a live video call with someone from your team, who hands over a single-use code once the candidate is visibly alone, and it is recorded. And it produces an item-by-item breakdown, so you can see not just a score but which kind of task was missed.

That last point is where the interview and the work sample meet. If the sample shows a candidate dropped every multi-step instruction, you now have something concrete to ask about in a follow-up conversation. "I noticed the four-step items were the ones you missed; how do you normally handle instructions with several parts?" That is a far better interview question than "how do you stay organised", because it is about something that actually happened.

Running the interview and the screen together

A sequence that works, from employers who use the Admin Skills Screen:

1. Interview first, as usual. Decide whether you like the candidate and whether the motivation is real. If not, stop; no screen needed.
2. Screen the shortlist. For the two or three candidates you would be willing to hire, run the ten-minute supervised exercise. The supervisor is on a video call, the candidate's camera and screen are recorded, and the result arrives the moment they finish.
3. Read the breakdown, not just the score. Look at which tasks were missed and whether the pattern matters for your role.
4. Follow up on specifics. If something surprised you, ask about it. Offer extended time if the candidate mentions a disability; the screen supports it with one setting.
5. Decide, with the interview, references and the screen in front of you. The screen is one input. You make the call.

For important roles, add a short paid trial task after the offer is accepted and before the start date. The screen tells you they can follow instructions; the trial tells you they will, with your actual work, over a week.

What this is not

A work sample does not measure intelligence, personality or potential, and a screen that claims to should be treated with suspicion. It does not replace judgement; it informs it. It should never be used as a cut-off that removes candidates without a human looking. And it should be the same for every candidate, under the same conditions, with the same accommodations available. Those principles are what make a screening step fair, and they are also what keep it defensible if anyone ever asks how a hiring decision was made.

Better interview questions, since you are asking anyway

If the interview is going to happen, it may as well be as informative as a conversation can be. Three changes help.

Ask for specifics, not categories. Instead of "how do you stay organised", ask "walk me through what you did with the last instruction you received that had more than three parts". Specific recall is harder to script and tells you whether the candidate actually notices multi-step instructions as a category.

Give one instruction in the interview and see what happens. "Before we finish, could you send me a short email summarising the three things we agreed, with the one that is time-sensitive first?" It is a tiny work sample inside the conversation. What arrives, and in what order, is informative.

Ask about the last mistake that reached a client. Not the safe, learned-from mistake; the one that got out. How they describe it, whether they describe the check that would have caught it, and whether they now do that check, tells you something the standard question does not.

None of these turn the interview into a reliable predictor. They nudge it from pure impression towards a little evidence, and they make the formal work sample afterwards feel like a natural continuation rather than an exam.

Where the interview still wins

It is worth saying what a work sample cannot do, because the argument here is for balance, not replacement. The interview is the only place you learn why the candidate wants this role, how they talk about previous employers, whether they ask good questions about your business, and whether you would be glad to see their name in your inbox every morning. Those things matter for a long working relationship, and no ten-minute exercise measures them. Keep the interview for them, and stop using it to guess at skills it was never able to see.

A closing observation

Ask any experienced owner about the best assistant they ever hired, and you will usually hear that the interview was unremarkable. Quiet, maybe a little nervous, not the most polished answers. What made the hire was that every instruction, from the first day, came back done exactly as asked. That quality does not show in a conversation. It shows in ten minutes of doing the work. The interview finds the people you would like to hire. The work sample finds the ones you should.

What changes when you add the work sample

Employers who add a ten-minute screen after the interview report the same shift: the second interview gets better. Instead of generic questions, the conversation turns to something that actually happened ten minutes earlier on a shared screen. "You caught the wrong day in the sentence straight away; what do you check before an invitation goes out?" "The four-step item with the skip was the one you missed; how would you handle that instruction on the job?" Candidates respond to these with real detail, because there is something real to respond to, and the employer learns in five minutes what forty minutes of prepared answers could not reveal. The interview stops being a performance and becomes a conversation about work, which is what it was always supposed to be.

Frequently asked questions

Won't candidates find a test off-putting after a good interview? Frame it as what it is: a short exercise so both sides can see the fit, recorded, supervised, the same for everyone. Candidates who are confident in their skills tend to welcome it.

Should I run the screen before the interview to save time? We advise against it. Used as a first filter it becomes a cut-off, which is both less fair and less informative. Interview first, screen the people you are serious about.

What if the screen contradicts a strong interview? Then you have learned something. Look at the item breakdown, ask about it, and weigh it. Sometimes a strong communicator is also careless with details; that is exactly the combination the interview cannot see and the screen can.