Skills assessment tools for hiring designers

Why design is badly served by the assessment market

Skills testing works best where an answer can be checked. Code compiles or it does not; a SQL query returns the right rows or it does not; a language test has a key. That is why the strongest, deepest libraries in this market are technical, and why they are genuinely good at what they do.

Design has very little of that surface. There is no correct number of steps in a checkout flow, no verifiable answer to whether a settings page should be tabs or a list. The decisions that separate a strong product designer from an adequate one are judgements made under constraints that change, and a test with a key cannot reach them.

So platforms do the reasonable thing and cover design with adjacent proxies — a visual-reasoning test, a tools test on Figma, a personality profile. Each measures something real. None measures the thing that made your best designer good.

The consequence is that most design hiring falls back on a portfolio and a conversation, which is not irrational: for craft, a portfolio is strong evidence and better than anything a platform will sell you. The gap is everything a portfolio structurally cannot show.

What a portfolio cannot show you, and why it matters

A case study is written after the fact by the person being assessed. That is not an accusation — it is the format. Everyone presents the version where the decisions were sound and the outcome was good.

What is missing is the part that predicts how they will work for you: what the brief said before they fixed it, which idea they argued for and lost, what they cut when six weeks of engineering disappeared, and whether they held a position when a founder pushed back twice.

Those are the expensive behaviours. A designer who builds exactly what was asked for, beautifully, can burn a quarter on a feature that was never going to move anything — and their portfolio will look excellent.

The five categories, and what each is worth for a designer

Ranked by how much they tell you about a design hire specifically, not by how good the products are in general.

• Portfolio review — the strongest evidence available for craft, taste and storytelling, and the fastest signal you will get. No platform substitutes for it. Blind to who decided what.
• Work samples and simulations (Vervoe, Canditech) — real evidence of output, and the method with the best general standing in selection research. The friction for design is cost: a representative design task takes hours, so in practice it gets used on two or three finalists rather than the pool where the decision is hardest.
• Situational judgment for the role — measures the decisions a portfolio cannot show, at a cost small enough to send to everyone. Thin ground commercially: few instruments are built for design specifically, which is why this category is usually missing from comparison articles. uxerhub is in it, and see the disclosure below.
• Broad test libraries (TestGorilla, Testlify) — worth buying if you also hire outside design, which most companies do. For the design hire itself, expect visual reasoning, tools knowledge and personality rather than design judgment.
• Psychometrics (Alva Labs, Bryq, Criteria) — the best-validated instruments on this list by a distance, and applicable to every role in the company. The limit for design is that two designers can return near-identical trait profiles and be right for opposite jobs, because the difference is in the match rather than the person.

A process that works, in the order that costs least

Cheap and wide first, expensive and narrow last. Almost every bad design hiring process inverts this and pays for it in candidate drop-off.

• Portfolio skim, five minutes. Filters for craft floor. Do not over-read it — you are checking whether the work clears a bar, not ranking six people who all clear it.
• One short judgment instrument, sent to everyone who clears the skim. Fifteen to thirty minutes of candidate time is an acceptable ask at this stage; ninety minutes is not, and the people who decline a ninety-minute unpaid task are disproportionately the ones with competing offers.
• A structured interview, same questions in the same order for everyone. This is the highest-return change most teams can make and it costs nothing. An unstructured conversation with six candidates produces six impressions that cannot be compared.
• A paid work session with the final one or two, if the role justifies it. Paid, because at this point you are asking for real work and there is a real prospect of an offer.

The questions worth asking any vendor, including us

Four that change the answer, and that no sales call volunteers.

• What has this been validated against, and on how many people? A straight answer names a criterion and a sample size. A logo wall is not an answer. Ours: scoring is deterministic arithmetic, and the instrument is not yet calibrated against hiring outcomes.
• How much unpaid candidate time does it cost, multiplied by everyone you assess? Longer is not more rigorous.
• What happens when a candidate contests the result? If a language model produced the score, explaining it is genuinely hard.
• What does it cost to run one real candidate this week, without a call? A free trial that requires a demo is a lead form.

Where we fit, stated plainly

uxerhub is in the third category and sells against several of the products named above, so read this section knowing that.

What it does: twelve written situations, about fifteen minutes, for six product roles including product designer. No option is correct, and the score is weighted by up to three capabilities you name as essential for the specific role before you send the link.

What it does not do: measure craft. It will not tell you whether someone can draw, and a portfolio review remains necessary. It also covers six roles only, is English-only, has no ATS integration, and is not calibrated — we have not run enough hires through it to publish predictive validity.

If you hire across many functions and can only buy one thing, buy a library. The narrow instrument is a supplement.

Frequently asked

What is the best skills assessment tool for hiring designers?

There is no single one, and any article that names one without qualification is selling something. For craft, a portfolio review outperforms anything on the market. For the decisions a portfolio cannot show, you need a judgment instrument. Most teams need both, and which specific products fit depends on whether you also hire outside design.

Can TestGorilla or Testlify assess designers?

Both have design-adjacent tests in their libraries — visual reasoning, tools knowledge, personality. They are useful and they are not design judgment. If you hire across several functions, a broad library is still worth buying for everything else.

Should I use a design take-home assignment instead?

A work sample is strong evidence, and the objection is cost rather than accuracy. A representative design task takes hours of unpaid candidate time, so it belongs at the end of a process on a small number of finalists, not at the top where the pool is large.

Do personality tests work for hiring designers?

They measure something real and well validated, but traits are stable and general while the design hire you need is specific. Two designers with near-identical profiles can be right for opposite jobs — one who questions every brief is what a founder needs and can be exhausting on a mature platform team.

How long should a design assessment take?

Short enough to send to everyone who clears your first filter, which in practice means under half an hour. Anything longer becomes a shortlist-only instrument, and the shortlist is not where the hard decision is.