Pre-employment assessment tools usually go wrong for a hiring team in one place: a score lands on every candidate’s applicant tracking system (ATS) record before anyone has agreed what it means. Each hiring manager then reads it their own way, some as a hard bar and some not at all, and the candidates who gave up time for the test hear nothing about what it was for.

The tools offer something interviews struggle with, because they set every candidate the same task under the same conditions. That only helps a team that knows which question the result answers and what a strong or weak result will change.

My view is that you choose the question before the tool. I wouldn’t buy or switch on any assessment until the hiring manager can say what the interviews for the role can’t show, and what they’ll do differently with the result. So I’ve set each kind of tool beside the question it can answer, along with a way to choose for one role at a time.

What pre-employment assessment tools measure.

Each kind of tool sets the candidate a different task, and the task decides what the result can tell you. Vendors bundle several kinds together, so it helps to separate them before you look at any one product.

Background checks and reference checks often arrive in the same bundle. They confirm a candidate’s history rather than measure what the candidate can do. They usually run close to the offer, under rules your legal or people team sets.

Why assessments earn a place, and what they cost.

Two interviewers asking the same opening question will follow up differently, so two candidates for one role rarely get tested on exactly the same thing. An assessment gives every candidate the same task under the same marking, which gives you evidence you can compare straight across a shortlist.

It also shows a skill directly, at a point when a resume shows less than it used to. With an AI writing assistant anyone can produce a polished application, and a polished application says nothing about whether the person behind it can build the model or handle the call. A short task the candidate does can settle that before anyone spends an hour of interview time on it.

The costs are just as real, and they land on candidates first. Every assessment asks for time, often unpaid and outside working hours, from people who have other offers to weigh. A test that doesn’t relate to the job can screen out people who would have done the job well, and a test that screens out more people from some groups than others can create legal exposure for the employer. A score also looks more objective than it is, so hiring managers tend to lean on it harder than the test was built to carry.

So an assessment has to earn its place again in every process it joins, one role at a time, and I’d test that before a candidate ever sees it.

Start from the question your interviews leave open.

Before you look at a single vendor, list what your interview loop for the role already covers, then write down the one thing it can’t show. That one thing is the question the assessment answers. Most gaps fall into a few shapes:

• A hands-on skill that takes longer to show than an interview allows, such as building a financial model or writing a long piece of copy.
• A skill no one on the panel can judge, such as fluency in a language your interviewers don’t speak.
• The same basic check across hundreds of applicants for a high-volume role, where interviewing everyone isn’t possible.
• Judgment in situations that rarely come up in an interview, such as a support agent deciding when to escalate.

Then match the kind of tool to the gap. For the support role, the question might be “Can this person write a clear, calm reply to an angry customer in the team’s tone?” A work sample answers that directly. A personality questionnaire tells you how the candidate describes themselves, which is a different question, however well the vendor’s report is designed.

If you can’t write the question down, don’t add a test. A role whose interviews already cover what the hiring manager needs is better served by a shorter process than by an extra score.

Agree what a result will change before anyone sits the test.

Settle with the hiring manager what happens after a strong result and after a weak one. There are only a few honest answers:

• The result decides who moves to the next stage, which makes the test a gate, and every gate needs the extra scrutiny in the vendor section below.
• The result shapes the next interview, so the interviewer probes whatever came back weakest.
• The result goes into the debrief beside the interview evidence, where the hiring team weighs it.

Write the answer into the role’s hiring plan, along with who sees the result and when. I’d share the candidate’s work with interviewers ahead of time, so they can ask about it, and keep the raw score for the debrief, so it doesn’t color how they rate the interview. If the answer to “what will it change?” is nothing, drop the test for this role.

Put the same questions to every vendor.

Vendor demos are built to look alike, and the differences show up in the answers to plain questions. Ask every shortlisted vendor the same ones, in writing, and compare the replies side by side.

Take the answers on pass rates to your legal or people team before launch, since testing rules vary from place to place. And a vendor that answers the first two rows with a brochure instead of documents has told you something about the test as well.

Pilot the test on people whose work you know.

Before a candidate sees the test, ask a handful of people already doing the role to take it, including some you’d call strong and some still finding their feet. You know their work. If their results don’t line up with what you see from them every week, the test is measuring something other than the job, and you’ve found out before it cost a candidate anything.

The pilot also tells you where the test belongs in the process. A short skills check can run early, when a role draws far more applicants than you can interview. A work sample that takes hours belongs after a first conversation, once the candidate knows enough about the role to judge whether it deserves the time. Tell candidates up front what the test is for and how the result will be used.

Take the result into the next interview.

In the next interview, ask the candidate to walk through the work sample: why they structured it that way, and what they’d change with another hour. Where one part of a skills test came back weak, ask a question that tests the same skill another way, because a weak section can come from nerves or an unfamiliar format.

Then have each interviewer rate the interview evidence on the scorecard on its own terms, and bring the test result to the debrief. The hiring team can then see where the two agree and where they pull apart, which is usually the most useful thing to discuss about a finalist.

For the probes themselves, this guide to follow-up interview questions names the gaps a first answer tends to leave and the follow-up that closes each one.

Where Metaview fits after the assessment.

Every part of the method above ends in an interview, where someone asks the candidate about their result and someone has to keep a record of the answer. The tests themselves come from your assessment vendor, and the interview that follows is where Metaview’s Notetaker comes in.

The Metaview Notetaker joins the interview as a visible participant. Where your team’s consent process has both people on the call agreeing to a recording, it records and transcribes the interview and sets the conversation out as structured notes, so the candidate’s explanation of their work sample is written down without the interviewer typing while they listen.

1

1. 1An answer held as three short notes under its “Q:” line, with the last note still being typed.

From the same conversation the Notetaker drafts a scorecard, which the interviewer reviews and rates before submitting. The draft is built from the candidate’s own answers, the evidence the debrief will set beside the test result.

Metaview captures every spoken word of a candidate’s walkthrough, which lets a debrief weighing a strong score against a shaky explanation read exactly how the candidate described their own work.

The case study on Raines International, an executive search firm, describes its process this way: “Candidate interviews are being transcribed and structured according to scorecards which are tied back to evaluation criteria.” For an assessment process, the lesson is to score the interview against the same criteria the test was chosen for, so the debrief compares evidence on one skill instead of two unrelated numbers.

In a debrief, a high score beside a weak walkthrough is a question for the room to work through, with the candidate’s own explanation open in front of everyone. I’d start with the walkthrough, because there the candidate explained their own work in their own words.

The record stays useful after the debrief. Type a question about a role into the Metaview Assistant and it answers from the interviews Metaview has captured, either summarizing what was said or recalling an exact detail, which helps when a role reopens and someone wants to know how the last shortlist handled the test.

Signs an assessment is no longer worth running.

An assessment that earned its place at launch can stop earning it as a role changes. A quarter after it goes live, put these questions to the hiring manager:

• Which decision on a recent candidate did the result change?
• Did anyone who scored well then struggle in the interview with the skill the test was meant to cover?
• Did candidates drop out of the process at the test itself?
• When the debrief opened, did people talk about the candidate’s work or about the number?

If the hiring manager can’t answer the first question, I’d take the test out of that role’s process, whatever the other three answers say, because a test that hasn’t changed a single decision in a quarter is costing candidates time for no return. Each of the other three sends you back to a step above, from the question you wrote to where the test sits.

Choosing a pre-employment assessment tool is that same review run before you buy: write down what the interviews can’t show you about the role, and pick the tool that answers it, with a result the team has agreed how to use.