153 rows answer "which test did I just get invited to", and 33 of them say no

Assessment vendors sell white-labelled products. The email a candidate receives comes from the employer, carries the employer's branding, and usually describes what they are about to sit as an online assessment. The most expensive preparation mistake is not practising too little; it is practising the wrong format for a week.

So one of our pages is a lookup table. The data is as dull as it gets:

{ terms: 'Verify, Verify G+, OPQ32, TalentCentral', label: 'SHL', href: '/games/shl' }

terms is what you might actually be holding: product names, aliases, misspellings, the URL in the invitation link, the title in the browser tab, the phrasing a careers page uses. label is what it is. href is where we can help.

153 rows. 147 point at a provider hub, 4 at one employer page, 1 at an article and 1 at a different product of ours. I counted them in the rendered DOM rather than the source, which turns out to matter, because the interesting thing about this table is not its size.

33 of the 153 labels disclaim something.

The six kinds of no

Eleven rows say the thing you searched for is retired, replaced or renamed:

{ terms: 'AT-SAT or ATSAT, the older FAA air traffic controller test',
  label: 'retired in November 2016. Practise the FAA ATSA instead',
  href: '/games/faa-atsa' }

{ terms: 'Postal Exam 473',
  label: 'A retired USPS exam; you will sit a Virtual Entry Assessment instead',
  href: '/games/usps' }

{ terms: 'Onetest', label: 'Revelian, renamed in 2014', href: '/games/revelian' }

Note where the bad news lives. It is in the label, which renders under the search terms in the row itself, not on the destination page. Somebody scanning for "AT-SAT" learns it is gone before they click.

One of those rows has no successor to practise. The Canadian Forces Aptitude Test was dropped in October 2024 and replaced with a scored application form, so its row reads replaced by the SEAF in October 2024. Not practised here and goes to an explainer article rather than a product.

Eleven more rows name something real and current that we deliberately do not simulate: aircrew batteries sat after a selection test, AI spoken interviews, assessor-marked written exercises, a US utility's three batteries. Each points at the section of a hub that explains why, because the useful answer to "we do not do this" is a paragraph, not a shrug.

Then there are the rows for things that are not tests at all. Applicant tracking systems, application portals, employer-side profiling tools:

{ terms: 'BrassRing, Talent Gateway, Infinite Talent',
  label: 'an applicant tracking system, not a test vendor. See the note below.',
  href: '/games/kenexa' }

{ terms: 'Ambertrack, or an application portal ending ambertrack.global',
  label: 'an application portal, not a test. Tests launched from it can be another vendor’s, so check the test link itself.',
  href: '/games/amberjack#ambertrack' }

The rows I like most are the ones that exist purely to send you away from the obvious match:

{ terms: 'USA Hire assessments for a TSA office or support job, such as Training Manager',
  label: 'USA Hire, not the TSA CBT',
  href: '/games/usa-hire' }

{ terms: 'Clevry, Criterion, Criterion Partnership, Psycruit, ... Leadership Judgement Test',
  label: 'Clevry (formerly Criterion Partnership). Not Criteria, which is a different company',
  href: '/games/clevry' }

{ terms: 'TMA, TMA Methode, TMA Method, Talenten Motivatie Analyse, ... TMA Performance',
  label: 'TMA, from TMA B.V. in Utrecht. Not Ixly, despite the shared city',
  href: '/games/tma' }

Nine of the 153 rows point at a fragment rather than a page, so the click lands on the paragraph that answers the question instead of the top of a long hub.

The rows are asserted, not maintained by goodwill

A table of 153 hand-written rows across 56 providers decays unless something holds it. What holds it is that each provider's own registration test owns its rows. The test locates the row by its destination and then checks the terms:

it('routes TSA CBT, TSA test and TAB to the TSA hub from the provider lookup table', () => {
  const source = read('app', 'blogs', 'which-assessment-provider-am-i-taking', 'page.tsx');
  const tsaRow = source.indexOf("href: '/games/tsa'");
  expect(tsaRow).toBeGreaterThan(-1);
  const rowText = source.slice(source.lastIndexOf('terms:', tsaRow), tsaRow);
  for (const term of ['TSA CBT', 'TSA test', 'TAB', 'TSO Assessment Battery']) expect(rowText, term).toContain(term);
  expect(source).toContain('TSA X-ray test');
  expect(source).toContain('USA Hire, not the TSA CBT');
});

Reading the page file as text rather than importing it is deliberate: these are server components inside a Next app, and the point of the assertion is the literal content, not a rendered tree.

The retired test has the stronger version of the same idea, a negative assertion:

expect(source).toContain("label: 'replaced by the SEAF in October 2024. Not practised here'");
expect(source).toContain(`href: '/blogs/cfat-replaced-by-seaf'`);
expect(source).not.toContain("href: '/games/cfat'");

That last line is the one that earns its place. The natural edit, a year from now, is for somebody to point the row at a provider hub that does not exist, because every other row in the table points at one.

See it

Open which assessment provider am I taking and count the cards in the grid under step 1. There are 153, and if you check their destinations you get 147 provider hubs, 4 to one employer page, 1 to an article and 1 to our interview product.

Then search the page for not. That is the third of the table that tells you we cannot help, or that the thing you typed is not what you think it is, or that it stopped existing. A lookup table for somebody else's product names is mostly an exercise in saying no accurately.

The same page says this about itself, which felt necessary: treat any employer to provider mapping you read online, including ours, as something to confirm against your own invitation. Employers change vendors between intake cycles, and last year's forum thread is weak evidence about this year's test.