Test Provider Guides

The publisher name in your invitation email is not the answer you think it is

Aon's own candidate preparation page lists 18 separate assessments it might send you. Two of them sit a few rows apart in the same table.

scales verbal gives you 49 tasks in 12 minutes. You read a dense block of text, then decide whether a series of statements is true, false, or cannot be determined from it. It is a reading and time-management exercise, and the pressure comes from volume.

motionChallenge gives you 6 minutes of interactive forward-planning puzzles. There is no reading passage at all. You are moving things around a screen and thinking several steps ahead.

Same publisher. Same brand on the invitation. Essentially no overlap in what you actually do. A candidate who spent a fortnight drilling verbal reasoning and then opens motionChallenge has prepared for the wrong test — and nothing in the email told them which one was coming.

That is the real problem with this landscape, and it is not the one most candidates worry about. The difficulty is rarely that any individual test is impossibly hard. It is that "you'll be taking an Aon assessment" narrows the field from thousands of possibilities down to eighteen, and eighteen is still too many to prepare for specifically.

The real test categories, and what actually separates them

Categories matter more than publisher names, because a numerical reasoning test from one publisher resembles a numerical reasoning test from another far more than it resembles a gamified working memory task from its own stablemate.

Numerical reasoning gives you tables, charts and financial data, and asks you to extract and compute. The thing candidates get wrong is preparing for the arithmetic rather than the extraction. In most of these tests, the calculation itself is straightforward — the difficulty is locating which of four columns you need and noticing that the figures are in thousands. Practising mental arithmetic helps less than practising reading a badly laid out table quickly.

Verbal reasoning is not a vocabulary test and not a comprehension test in the school sense. The dominant format asks you to judge statements as true, false, or cannot say — and the entire skill is disciplining yourself to answer only from the passage. Candidates who know a subject well perform worse on questions about it, because they answer from knowledge instead of from the text.

Inductive and logical reasoning uses shapes, sequences and patterns rather than words or numbers. Aon's scales cls, for example, gives 12 tasks in 12 minutes. What separates strong performers is method: isolating one rule at a time across a whole sequence rather than trying to perceive the pattern whole. Most sequences run two or three rules simultaneously at different frequencies, and looking for a single unifying pattern is the standard way to fail.

Situational judgement presents workplace scenarios and asks you to rank or choose responses. This is the category where "practice" means something different — you are not building a skill, you are learning what a particular employer values. Aon's chatAssess runs 20 minutes and delivers scenarios through a messaging interface rather than a question list.

Personality and behavioural questionnaires are where preparation is most often actively counterproductive. Aon's ADEPT-15 asks you to respond to 100 statements over roughly 30 minutes. Talogy's PAPI covers 7 factors across 26 scales and runs from 44 to 162 items depending on version. Instruments at that length include consistency checks by design. Trying to construct an ideal candidate across 100 statements tends to produce a profile that reads as internally contradictory, which is worse than an honest one.

Interactive and task-based formats are the newest category and the one least served by conventional practice. Criteria's Cognify runs three interactive games in about 10 minutes. Arctic Shores has moved away from questions entirely toward task-based assessment lasting around 45 minutes, measuring what it calls Skill-enablers. There is no question bank to work through, because there are no questions.

The single most common preparation mistake is spending all available time on numerical and verbal reasoning — the two categories with the most practice material available — when the assessment ahead is behavioural or interactive.

How to figure out which provider and format you're actually facing

Be realistic about what you can establish in advance. Some of it is genuinely knowable, and some of it is not.

What the invitation usually tells you. The link domain often reveals the publisher even when the email body doesn't name one. The email will normally state the number of assessments in the batch and the completion window — commonly a set number of days rather than a fixed appointment. Many invitations state the duration of each section, and that number alone narrows things considerably: something lasting 6 minutes and something lasting 30 minutes from the same publisher are almost certainly different categories of test.

What is worth asking your recruiter. Ask what the assessment covers rather than which product it is. "Is there a numerical component?" and "Is any part of it timed per question rather than overall?" get useful answers. "Which product is it, exactly?" often does not, because the recruiter may not know — assessments are frequently configured by a central talent team and administered by someone who only sees the invite-and-chase workflow. Asking about accommodations is always legitimate and usually gets a substantive reply.

What you will only learn once you start. Most publishers provide practice items or an untimed tutorial before the scored section begins, and these are the most valuable thing available to you. They are the only place you see the actual interface. Read every instruction rather than clicking through — whether you can revisit earlier questions, whether the test adapts to your answers, and whether there is a per-question limit are all things you learn there and cannot reliably learn beforehand.

Be sceptical of any guide claiming to know a specific employer's exact test configuration. Employers change publishers, publishers retire products, and configurations vary by role and region. A page confidently stating a time limit for a named employer is frequently describing something that was accurate three years ago.

Why the same provider name can mean wildly different tests

This is the part that catches out experienced candidates specifically, because prior exposure creates false confidence.

Criteria's own 2026 assessment portfolio lists 29 distinct products. Its CCAT is a conventional multiple-choice cognitive test running 15 minutes. Its Cognify is three interactive games running about 10 minutes. Both are Criteria cognitive assessments. Someone who has taken one has learned almost nothing transferable about the other beyond the company's logo.

Saville's Swift range alone contains six different combination tests. Swift Comprehension Aptitude runs between 9.5 and 16 minutes and covers verbal, numerical and error checking. Swift Technical Aptitude runs 10 minutes and covers spatial, mechanical and diagrammatic reasoning. Swift Global Aptitude runs 12 minutes and is highly visual, including memory recall. Three products, one family name, three genuinely different experiences.

Hogan is a cleaner illustration still. Three of its best-known instruments — the HPI, the HDS and the MVPI — each take 15 to 20 minutes and are all personality or values inventories with no right answers. Its HBRI takes around 35 minutes and is a reasoning test that does have right answers. Being told "you're doing a Hogan" tells you nothing about whether you are about to be scored on correctness at all.

There is a second complication worth knowing about: publisher names change. Cut-e was acquired by Aon and no longer exists as a live brand. Cubiks became part of Talogy in 2022, when 16 businesses were unified under one identity. Revelian was absorbed into Criteria, with its Cognify and Emotify products retained under the new name. Older guides — and older colleagues — will refer to products under brands that the publisher itself has retired, which makes secondhand advice about "the cut-e test" or "the Cubiks test" less reliable than it sounds.

How to prepare when you don't know the exact format yet

Early in a job search, before any invitation arrives, the efficient move is to build the things that transfer and skip the things that don't.

Worth doing early, because it transfers. Reading tables and charts quickly, including spotting units and footnotes. Percentage change, ratios and unit conversions without a calculator. Answering verbal items strictly from the passage. Isolating individual rules in a pattern sequence rather than looking for one grand pattern.

Worth doing early for a different reason. Sit one timed assessment of any kind, start to finish, before it matters. Most candidates have never experienced a countdown clock while reading a table, and discovering your reaction to that during a real assessment is expensive.

Worth waiting for. Publisher-specific interface familiarity, and situational judgement practice — employer-specific and largely wasted before you know who is asking.

Not worth doing at all. Rehearsing personality questionnaires. Memorising claimed pass marks from prep sites, since publishers essentially never publish benchmarks and the figures circulating online routinely contradict one another.

One mundane detail repays attention out of proportion to its difficulty: check the technical requirements the moment the invitation arrives rather than on the day. Some assessments are desktop only, and interactive and memory-based formats are materially harder on a small screen.

FAQs

Do all employers use the same test provider?

No, and many large employers use several at once for different roles or regions. Technology roles frequently sit on a coding platform while commercial roles at the same company use a conventional reasoning battery. It is entirely normal to face a completely different assessment applying to two teams within one organisation.

Can I ask my recruiter which test I'll be given?

Yes, and it is a reasonable question — but ask what it covers rather than which product it is. Recruiters can often confirm whether there is a numerical section or how long the batch takes, while the specific product configuration is frequently set centrally and not visible to them. Questions about accommodations should always be asked and are always legitimate.

How different can two tests from the same publisher be?

Very. Aon's own candidate page lists products ranging from an assessment of 2 minutes measuring concentration and reaction speed to a behavioural questionnaire of 100 statements taking around 30 minutes. Talogy lists 12 named products spanning cognitive ability, personality, emotional intelligence and leadership. Prior experience of one tells you the interface conventions and very little else.

Is it worth practising a format I might not even face?

Numerical and verbal reasoning, yes — they appear across most publishers and the underlying skills transfer. Publisher-specific interfaces and situational judgement, generally not, since both are wasted if the format turns out to be different. The exception is sitting any timed test once for the experience of the clock.

Why do older guides mention providers I can't find?

Because the industry consolidates constantly. Cut-e is now Aon, Cubiks is now Talogy, Revelian is now Criteria. Products get retired and renamed while guidance written about them stays online, so advice referencing a retired brand may describe an instrument that no longer exists in that form.

Should I prepare differently for a personality questionnaire?

Yes — mostly by not preparing. Instruments of this length include internal consistency checks, and answering according to a constructed ideal across 100 or more statements tends to produce a contradictory profile. Reading the instructions properly and answering at a steady pace is the whole strategy.