All guides

What a Virtual Assistant Skills Test Should Actually Measure

Most "VA tests" measure typing speed and vocabulary. The skills that decide whether a remote assistant works out are different, smaller and testable in ten minutes. Here is what to test and how.

Search for "virtual assistant skills test" and you will find typing tests, English grammar quizzes, personality inventories and software certifications. Each of those measures something. None of them measures the thing that actually decides whether a remote assistant will save you time or cost you time: whether they can take an instruction, hold it, and carry it out exactly, when nobody is sitting next to them.

This page sets out what a useful skills test for an admin or virtual-assistant role should measure, why the usual tests miss it, and how to run a screen that catches it in ten minutes.

Start from the job, not the test catalogue

The right way to design any screen is to list the small things the job consists of and work backwards. For a remote administrative role, that list is short and remarkably consistent across industries:

  1. Receive an instruction, usually once, usually verbally or in a chat message, often with more than one step.
  2. Carry it out in the right order, including any exception that applies this time.
  3. Keep track of names, dates, amounts and sequences for long enough to act on them.
  4. Produce output with no wrong dates, names or figures.
  5. Apply the simple rules of the business to each case without being reminded.
  6. Know what day it is, what is due, and what comes next.

Everything else, the specific software, the industry vocabulary, the tone of emails, can be taught to a person who can do those six things. None of it can be taught to a person who cannot.

So the test should sample those six things, directly, with tasks that look like the job.

What the usual tests measure instead

Typing speed. Useful for a transcriptionist. For an admin, a 40-words-per-minute typist who follows instructions exactly beats a 90-words-per-minute typist who misses the second step, every day of the week.

English tests. Vocabulary and grammar matter for client-facing writing. They say nothing about whether someone can hold a three-item list. Plenty of eloquent people cannot, and plenty of people with modest English are meticulous.

Personality questionnaires. They measure how a person describes themselves. "I am very detail-oriented" is a sentence anyone can produce. The only way to know whether someone is detail-oriented is to give them a detail and see whether they orient to it.

Software certifications. Proof that someone sat through a course. Helpful at the margin, irrelevant to the core.

Long unsupervised assessments. The candidate takes a forty-minute online test at home. You get a score. You do not know who took it, whether they had help, or whether they Googled the answers. The score is unverifiable, which makes it worthless as evidence.

The six things, as test items

Here is how each of the six job behaviours turns into a short task. The Admin Skills Screen uses tasks of exactly this shape, with the specific words and numbers drawn fresh for every attempt.

Following a multi-step instruction with a twist. "Do these four things in order, but do step two last." Then, later, a harder one: "do step one, skip step three, and insert this before what you already typed." A candidate who can hold the instruction and apply the twist will do the same with "process these as usual, except the international ones".

Remembering a short schedule after it disappears. Three day-and-task pairs appear for fifteen seconds, then vanish. Which task was on Wednesday? Which task came last in the week? This is the Monday-morning call where you rattle off the week and expect it to stick.

Correcting an error in text. A sentence with the wrong weekday. Retype it correctly. Simple, and the single most common real-world admin failure is the wrong date going out unnoticed.

Small arithmetic. Add two numbers. Sort three and pick the middle one. Not because admin work is mathematical, but because an assistant who cannot reliably add two figures will eventually confirm the wrong amount to a supplier.

Applying a rule to a case. "If your name has more than five letters, mark it underline, otherwise circle." A trivial rule, deliberately, because the skill being tested is whether the rule gets applied at all, not whether the candidate can handle a hard rule.

Knowing the day. Is today a weekday or a weekend, and what day is it? Nobody fails this through ignorance. People fail it by not reading the question, which is itself the finding.

Carrying something forward. Several items later: "what was the middle number you found earlier?" Because real work does not arrive one task at a time; it arrives interleaved, and the assistant has to hold a figure from the morning until the afternoon.

Thirteen short tasks of that kind, in about ten minutes, produce a score out of seventeen and an item-by-item breakdown. The breakdown is worth more than the score. Two candidates at sixty percent can have completely different profiles: one who is careful but slow with arithmetic, one who is fast but drops every multi-step instruction. For an admin role those are not the same hire.

What makes the result trustworthy

A test is only as good as the conditions it was taken under. Four conditions turn a quiz into evidence:

What a skills test should not try to do

It should not try to measure intelligence, personality or "fit". Those are either unmeasurable in ten minutes or none of a screening tool's business. It should not be long; beyond ten or fifteen minutes you are measuring stamina. It should not be adaptive or clever; the tasks are basic because the skills are basic, and an assistant who finds them tricky will find the job tricky. And it should not claim to predict a career. It predicts the first month, which is the month that costs you.

A worked example: two candidates, one score

Here is a pattern we see often enough to be worth describing. Two candidates for the same remote admin role both score 59 percent, which lands both in the Borderline band. On the number alone they are interchangeable. The breakdowns say otherwise.

Candidate A got every instruction-following item right, including the reorder with the skip, and every correction right. She lost points on arithmetic (a wrong sum under time pressure) and on one memory item where she recalled two of three bank names in the right order. Her profile: careful, exact, slightly weak on retention under pressure.

Candidate B got the arithmetic right, the day-of-week item right and the correction right, and missed every item that involved a twist or a carried-forward figure. His profile: fine in routine, loses the thread when the instruction departs from the expected pattern.

For a role with daily exceptions and verbal instructions, A is the far better bet and B is the profile that produces the "I told them exactly what to do" conversation three weeks in. For a role that is almost entirely routine with written procedures, B may be perfectly adequate. The score could not have told you that. The breakdown did, in a glance, and the follow-up questions in the second interview wrote themselves.

What to tell candidates in advance

Transparency improves results and fairness at the same time. A short note before the call should say: it is a ten-minute exercise of everyday admin tasks, taken on a video call with a supervisor present, recorded with camera and screen, one attempt, nothing to study. Candidates who need extra time for a disability should say so beforehand and it will be arranged. Candidates who know what is coming perform as they will perform on the job, which is what you want to see; surprise measures composure, not skill.

Using the test with existing staff

Employers often run the screen on their current team first. It is a low-stakes way to see what "good" looks like in their own context, to confirm the exercise feels fair before candidates take it, and to set a benchmark. It is also, quietly, informative: more than one owner has discovered that the assistant they trust most scored highest, and the one they spend the most time checking scored lowest, which is the test doing precisely what it is for.

How the result fits into the rest of the decision

A skills test is strongest when it is the one piece of hard evidence in a decision that otherwise runs on judgement, and weakest when it is asked to be the whole decision. Employers who get the most from it treat the breakdown as a map of where to look. A clean breakdown means the first month can start with trust: give instructions once and expect them carried out. A breakdown with one weak category means a specific habit to install from day one, written read-backs or a checklist, and a review point at the end of week two. A breakdown with several weak categories, after a strong interview, is the signal that the interview was measuring presentation, and the decision should weigh that honestly rather than explain it away. In each case the test did not decide; it told a person what they were deciding about.

A short checklist for choosing any skills test

If you are evaluating this screen or any other, ask:

A test that answers yes to all seven is a work sample you can defend and rely on. A test that answers no to the first one is measuring something other than the job, however sophisticated it looks.

Frequently asked questions

Is a ten-minute test really enough to judge someone? It is enough to see whether they can follow instructions and hold details, which is what it claims. It is not enough to judge the whole person, which is why it is one input and the interview still happens.

Will good candidates be put off by being tested? In our experience, no. Capable people like a chance to show it. The candidates who object most strongly to a short, transparent work sample are, more often than not, the ones the sample would have caught.

Can I see the questions before I use it? You can take it yourself in ten minutes, which is the better way to understand it. The specific values you see will never appear for a candidate.

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 →