All guides

Why You Should Test a Virtual Assistant's Basic Skills Before You Hire

A founder's account of what an untested admin hire actually costs, and why a ten-minute supervised skills screen before the offer beats months of fixing mistakes after it.

I learned this the expensive way. I hired a remote administrative assistant on the strength of a good interview, a pleasant manner and a CV that said all the right things. Within a month I was spending more time correcting her work than the work itself would have taken me. A calendar invitation sent for the wrong week. A supplier told the wrong amount. A three-step instruction, given once over a call, that came back with step two missing. None of it was malicious. She simply did not have the basic skills the job needed, and nothing in my hiring process had ever checked.

That is the whole argument of this page, and of the product behind it. The skills that make or break an admin hire are small, concrete and testable in ten minutes. Almost nobody tests them. Everybody pays for not testing them.

The skills that matter are not on the CV

When people talk about "admin skills" they usually mean software: Excel, a CRM, a calendar tool. Those can be taught in an afternoon. The skills that cannot be taught quickly, and that decide whether a remote assistant is an asset or a liability, are more basic than that:

None of those appear on a CV, because nobody writes "I sometimes miss the second step" on a CV. None of them come out in an interview either, because an interview is a conversation, and a conversation rewards people who are good at conversations. The candidate who talks fluently about "attention to detail" may or may not have any. You find out in week three, when the detail is a client's invoice.

Why it is so much cheaper to find out before

Here is the arithmetic that convinced me, and it is not complicated.

A bad admin hire does not fail loudly. A bad engineer ships a bug and the system tells you. A bad admin sends the wrong date, and nothing tells you until the client turns up on the wrong day. The cost of each mistake lands somewhere else in the business, often on you, and often weeks later. By the time the pattern is clear, you have paid for one to three months of salary, lost the time you spent checking and redoing work, lost some goodwill with whoever was on the receiving end of the mistakes, and you are back at the start of the hiring process. If you are honest about your own hourly rate, the cost of one wrong hire is rarely less than three months of that person's pay and is often much more.

Against that, a skills screen before the offer costs ten minutes of a candidate's time and ten minutes of a supervisor's. The candidate who cannot hold a three-item list or follow a reordered instruction does not get hired, and nothing downstream ever happens. There is no version of the maths where finding out later is cheaper than finding out first.

What a good screen looks like

Not every test is worth running. A long personality questionnaire tells you how a person describes themselves. An unsupervised online quiz tells you how good their friend is at quizzes. The screen that actually predicts the first month on the job has four properties:

  1. It is a work sample. The tasks on it are the tasks of the job, in plain business language: an invoice, a delivery day, a bank name, a weekday. Not abstract puzzles, not trick questions.
  2. It is short. Ten minutes is enough to see whether someone can follow instructions and hold details. Forty minutes mostly measures stamina and goodwill.
  3. It is supervised and recorded. Someone from your team is on a video call for the whole thing, hands over the single-use code live, and the candidate's camera and screen are recorded. A result you cannot trust is worth nothing.
  4. It is one input, not a verdict. A score and a plain description go to the person making the decision. The software never accepts or rejects anyone. That is both the ethical position and the legally defensible one.

The Admin Skills Screen was built to those four properties, after the hire described at the top of this page. It is a ten-minute exercise, freshly randomised for every attempt so there is nothing to pass around, taken on a live call with your supervisor present, recorded from start to finish, and scored automatically against the exact values the candidate was shown.

What it is not

It is worth being precise, because the hiring world is full of products that over-claim. This is not an intelligence test and it does not measure "aptitude" in the sense a psychologist would use the word. It does not predict long-term career success, and it has not been through a validation study that would let anyone claim it does. It is a practical check of whether a person can do, today, the small concrete things a remote admin role requires on day one. Treat it as you would treat watching someone type for ten minutes before hiring a typist: not the whole decision, but a piece of it you would be foolish to skip.

The remote twist

If you hire in the same office, you get a hundred small signals for free. You see whether someone writes things down. You see whether they ask when they are unsure. You hear them repeat an instruction back. Remote hiring strips all of that away, and offshore hiring strips it away across a time zone and a cultural distance as well. The interview becomes the only signal, and the interview is the weakest signal there is.

That is why the screen matters more, not less, for remote and offshore roles. It puts back one concrete, observed, recorded piece of evidence into a process that otherwise runs on impressions. Our own experience, and that of every employer we have spoken to since, is that the ten minutes on the screening call tell you more than the forty-minute interview that preceded them.

The three mistakes that told me everything

Looking back, the three mistakes that made me finally act were not dramatic. That is the point. They were ordinary, and they were the kind of thing the job consists of.

The first was a client meeting booked for the right day of the wrong week. The instruction had been "Thursday the 14th". The invitation went out for Thursday the 7th. The client turned up at an empty video room, waited, and emailed me directly, politely, to ask whether we had forgotten. A thirty-second slip became a week of me checking every invitation personally.

The second was an amount. A supplier had quoted two figures on a call, one for the standard order and one with the rush fee. I asked for the standard one to be confirmed in writing. The rush figure was confirmed instead, because the two numbers had been heard a minute apart and only one of them had stuck. The supplier, reasonably, held us to it.

The third was the one that made me stop and think. I gave a four-step instruction over a call, clearly, and asked for it to be repeated back. It was repeated back correctly. Then it was carried out with step two missing. Nothing was wrong with the listening. Something was wrong with holding four things and acting on them in order a few minutes later, and that is a skill, not an attitude, and it is one I had never checked.

Every item on the screen we built afterwards maps onto a mistake like those. Not by coincidence: we wrote the items by listing the mistakes.

How to use it well

A few things we have learned from running it on our own candidates:

A word about offshore hiring specifically

Much of our own hiring is offshore, and much of the interest in this screen comes from employers who hire the same way, so it is worth being direct about what is different and what is not.

What is not different: the skills. A remote assistant in Manila, Lagos or Lisbon needs exactly the same ability to follow a multi-step instruction and hold a schedule as one in the next room. The screen does not know or care where the candidate is.

What is different: the information you have. Offshore labour markets are intensely competitive, which is what makes them affordable, and intense competition pushes every candidate to present the best possible version of themselves. Experience is rounded up. Agencies describe everyone as "highly skilled". A polished profile and a warm interview are the entry ticket, not a differentiator. Meanwhile you are hiring across a time zone and often across a language, so the informal signals that would warn you early are weaker and arrive later. The result is that the gap between how a candidate presents and how they perform is wider on average, not because of who they are but because of how the market works and how little you can see.

A supervised, recorded, ten-minute work sample is the cheapest way to close that gap. It puts one piece of direct evidence into a process that otherwise has none, and it does so in a way that is the same for every candidate, wherever they sit.

What happened next

After the hire that started all this, we tested the next three candidates for the same role with an early version of this screen. One of them scored far below the others despite the best interview of the three. We did not hire her, and I will never know whether that was the right call, which is the honest truth about any screening decision. The one we did hire followed every instruction on the screen exactly, has followed every instruction since, and has made exactly one calendar mistake in a year, which she caught herself before it mattered.

That is what a ten-minute screen buys you. Not certainty. A much better bet, before the money is spent instead of after.

Frequently asked questions

Is this fair to candidates? Fairer than an interview, in our view. Everyone gets the same tasks under the same conditions, the tasks are the job, and a human with the context decides. The accommodations exist for those who need them. The alternative, hiring on impression, is the least fair method there is.

Can candidates prepare for it? They can know the format, and we tell them. They cannot know the content, because every attempt draws fresh values from pools. Knowing that you will be asked to remember three bank names does not help you remember three bank names.

How long until I see the result? The moment the candidate finishes. The score, the band, the item-by-item breakdown and both recordings appear on your dashboard.

Does it replace the interview? No. It replaces the part of the interview where you were guessing.

The Admin Skills Screen

A supervised, recorded, ten-minute screening exercise for remote admin hires. Single-use codes, fresh content every attempt, one input to your decision.

See how it works →