Outmatch Assessment 2026: Four Match Bands, No Cutoff

The name on your invitation, and the company that owns it now

Start with the name, because everything else hangs on reading the right documentation. Outmatch acquired Harver in May 2021 and then took the acquired company's name. The rebrand announcement is dated 17 November 2021 and filed from Dallas, and it says the merged firm "rebrands to Harver and continues to develop the world's leading volume hiring solution for hourly and professional roles", checked 7 October 2026.

Typing the old address proves the point in one request. The outmatch.com domain answers with a 301 redirect to harver.com, so the name the market still types no longer serves a site.

Three pieces of the old name are still in service, which is why candidates get confused. The candidate help center lives at support.outmatch.com and is branded Outmatch. System mail leaves from a noreply address on outmatchapps.com. And the consent article tells administrators that "the client's privacy policy does not supersede Harver's policy", so Harver's terms govern inside a product labeled Outmatch.

For a candidate this is practical rather than trivial. The corporate pages on harver.com describe what the product does for an employer. The help center on support.outmatch.com describes what happens to you. Both are the publisher. Only one of them answers your questions.

What an Outmatch assessment contains, module by module

The invitation decides which modules you get, and the publisher documents each one separately. Four appear most often, and the durations below are the publisher's own.

The cognitive page names its five subtests: logical reasoning, verbal reasoning, perceptual speed and accuracy, numerical reasoning, and spatial ability. It also states that the set is yours to configure: "Because of the modular set-up of the assessment, you can determine what you want to measure per role and vacancy", checked 7 October 2026.

One honest caution about that page. Its body lists the five subtests above. Its own FAQ, a few hundred words lower, lists "numerical reasoning, verbal reasoning, logical reasoning, spatial ability, and working memory". Working memory appears only in the FAQ and perceptual speed only in the body. The publisher contradicts itself about its own composition, so no page should promise you a fixed set. Pace the section you are given, the way an SHL sitting is paced by seconds per item.

The personality questionnaire rests on HEXACO and comes in two lengths. The situational judgment test asks you to pick the best and the worst option in each scene, which is a different task from picking one answer.

The four match bands, and the jump the publisher calls rare

This is the part nobody quotes, and it sits in an administrator article about invitation contexts. The publisher names four result bands: Poor Match, Fair Match, Strong Match and Strongest Match.

The sentence that names them is about movement between them. "Regardless of the context selected, we don't typically see a large jump from a Poor Match to a Strong or Strongest Match. It is much more likely the shift would be from Poor Match to Fair Match", checked 7 October 2026 in the assessment context article.

Read what that concedes. The bands are stable across the two ways the publisher can administer the same instrument, and a weak result tends to stay weak when the conditions change. The instrument is doing something consistent, and no amount of resitting is designed to move you two bands.

The recruiter sees more than a label. A March 2024 report redesign added a "Band Label" so a reader "can understand where the candidate overall score lies within the range", and it shows a percentile against the median of the comparison group. The same note says the weight carried by the personality part against the cognitive part is configurable per account, and that where only personality is present "its relevance is 100%".

So there are four named bands, one percentile, and a split between the two halves of the score that changes from employer to employer. HireVue publishes numeric ranges for its bands. This publisher names its bands and publishes no ranges at all.

Selection or Development: the context that moves your norms

Every invitation carries a context, and the choice changes the norm your answers are compared against. The publisher documents both.

Selection covers external candidates applying for a job, internal employees applying for a promotion, and succession planning. Development covers self-insight, coaching and workshops. As a job applicant you are almost always in the first one. The article states that "all jobs set up through an ATS partner are in the Selection Context by default", checked 7 October 2026.

The reason for two norms is behavioral, and the publisher spells it out. "Individuals who complete the assessment in a Development context tend to respond more candidly than applicants who complete in a Selection context." Applicants, it says, tend to put their best foot forward and manage the impression they leave. People scored for development therefore come out lower on overall fit, and separate norms were built so the two populations are not compared with one ruler.

Two consequences follow for a candidate, and both are useful. Everyone in your comparison group was also presenting themselves at their best, so presenting well is the floor rather than an edge. And a development report from a previous employer tells you very little about how you would score as an applicant, because it was read against the other norm.

If an invitation does not say which context applies, the person to ask is the recruiter. The publisher's own candidate article routes timing and process questions to the employer and keeps only technical faults for itself.

The video interview, screen by screen

The recorded interview is the module candidates worry about, and the publisher documents it step by step. The walkthrough article describes the order of screens, checked 7 October 2026.

1. A welcome video and a hints video play first. They explain the mechanics. Nothing is recorded yet.
2. A practice question comes next, with no limit on retries. In the publisher's words, "you can re-record yourself as many times as you need until you feel ready to continue to the timed interview questions."
3. Each real question appears on screen with a short preparation window. The publisher puts it at "usually between 30 and 60 seconds, but can vary", so the number on your screen is the only one that counts.
4. Recording starts by itself. A countdown runs in the top right, and a Stop button submits early when you have finished.
5. An upload confirmation appears between questions. You click to continue.

Two settings are worth knowing before you start. A Blur my Video option hides your own image from yourself, and the publisher notes that it "will blur the video on your screen only", so the reviewer still sees you clearly. And the whole sitting runs "between 15 and 25 minutes" with the advice to "set an hour aside", which leaves room for a hardware problem without losing the window.

The rule that costs people a candidacy is about leaving. Before the real questions, you can close the tab and come back. After them, the publisher is blunt: "seeing a question and leaving the process will mean that you can no longer answer that question." Treat the practice screen as the last exit.

Where an Outmatch assessment is sold for consulting hiring

The publisher sells this product into consulting hiring by name, and that is the only published link between the tool and our market. Its professional services page says "whether you need accountants, recruiters, or consultants, Harver's assessments reliably determine fit in as little as 16 minutes for professional jobs", and elsewhere on the same page it describes "building teams in consulting, finance, legal, or business support", checked 7 October 2026.

Hold that 16 minute figure against the module table above. A single cognitive section runs 6 to 24 minutes by the publisher's own FAQ, and a video interview runs 15 to 25. A 16 minute sitting is therefore a narrow configuration, probably one short section and a questionnaire. Any consulting employer who adds a video interview has already left that number behind.

Now the part the rest of the internet skips. The same page advertises a 40 percent fall in time to hire, a 45 percent fall in turnover and a 97 percent candidate satisfaction score, and it publishes no method, no sample and no period for any of the three. They are claims by a seller about its own product.

And no consulting employer is named anywhere. The rebrand announcement prints its client list in full, and it reads "category leaders such as Sitel Group, Valvoline, McDonald's, Chili's, and more", which is volume and hourly hiring. The situational judgment page says those tests are "especially effective for any that require volume hiring, like retail client services and supply chain and logistics". We looked for a consulting firm that names this vendor on a careers page and found none.

So the honest statement is narrow. The vendor sells this for hiring consultants and says so in its own words. Which firm actually sends it, for which role, in 2026, is in your invitation email and nowhere else. For the exercises firms build themselves instead of buying, see consulting work simulations.

One sitting, four other jobs, no second action from you

Here is a mechanic with no equivalent at the other screening vendors. The publisher's Succeed module replays the answers you already gave against the scoring profiles of other roles in the same account.

The article is explicit about the cost to you, which is nothing. A Succeed report carries the full overview, insights, interview guide and development sections "aligned with the selected Job profile, all without any action by the candidate/employee", and a recruiter "can select up to 4 jobs to compare", checked 7 October 2026 in the Succeed article.

Two limits are published alongside it. Only jobs built on the same underlying questions appear in the picker, so the replay works within a family of roles rather than across the whole company. And "Succeed reports do not include Ability Measure results", which means a rescore against another job uses the behavioral and competency parts only.

The practical reading is calm rather than alarming. One sitting can put you in front of more than one opening without a second invitation, and the parts of your result that travel are the personality and competency parts. Nothing in the documentation says a candidate is told when this happens.

Three ways a sitting goes wrong, and who can fix each one

The publisher divides responsibility clearly, and knowing the division saves a day of mail to the wrong address.

1. You left after the practice question. This is the self-inflicted one, and the published remedy is a retake you cannot grant yourself. The employer chooses which questions to reopen, sets a new deadline and attaches a message, per the candidate management article. So the request goes to the recruiter.
2. The connection dropped mid recording. The publisher's advice is to contact it or the employer, and it notes that the assessment "doesn't require a fast connection, but it does have to be stable". A cable beats wireless for twenty minutes.
3. The platform threw an error. Here the publisher owns the problem, and it asks for your name, the company you are applying to, the page you were on, and your browser and device. That it needs to ask which company tells you something: the vendor does not know your employer from the link alone.

One blocking screen catches people before any of this. When an employer turns on participant consent, a checkbox gates the assessment, its wording is fixed by the publisher, and nothing proceeds until it is ticked. The feature is off by default, so it appears for some employers and not others.

Accessibility is documented and partial. High contrast mode, screen reader text and keyboard focus indicators are supported, and the publisher states that "there are some aspects of Assessments that cannot feasibly be made compliant, such as complex charts and graphs." Anyone who needs an adjustment should ask the employer before the window opens.

What no page publishes about an Outmatch assessment

Five gaps, and naming them is more useful than filling them with guesses.

The cutoff. Four bands are named and no decision is attached to any of them. Nothing in the publisher's documentation says which band advances, and the employer sets the bar per role.

The weighting. The split of relevance between the personality half and the cognitive half is a per account setting by the publisher's own note, so there is no single number to publish. A page that tells you personality counts for a fixed share of your result has invented it.

Your score. The documentation describes the report built for the recruiter, in detail, and says nothing about a result shown to the candidate. The personality page offers employers the option of sharing results afterwards, which makes it a choice rather than a right.

Whether a human looks first. No article says a person reviews before a rejection, and none says the opposite. We decline to assert either.

Which consulting employer uses it. Covered above. The vendor publishes a client list and no consulting firm is on it, the vendor sells the product for hiring consultants, and the two facts coexist without resolving.

On access, the method is part of the source. The help center is behind Cloudflare and returns 403 to a normal fetch, so the 86 articles were read through the public Zendesk API on the same host. The corporate pages answered directly with a browser user agent. Nothing above rests on a page we could not open.

Four moves before the window opens, and what our tools leave alone

Four moves, in the order they pay off.

1. Read the invitation for the module list and the clock. The modules vary, the preparation window varies, and the screen in front of you outranks every range on this page.
2. Spend the practice question properly. It is unlimited and unscored. Say one full answer, play it back, and fix the microphone before anything counts.
3. Build six short stories with a result in the last sentence. The recorded questions reward a situation, your decision and a measured outcome, in that order.
4. Finish in one sitting once the real questions start. The published rule about abandoning a seen question is the single most expensive thing on this page.

Our free drills train the part that carries over, which is stating a conclusion first and supporting it in two moves against a clock. How an answer is scored publishes the weights behind them, read from the code that does the scoring, and those weights are ours. The publisher names bands and no rubric, and no employer publishes a bar. Our assessment simulators cover formats firms build themselves, and that page says plainly that none of the six is live yet. For the publishers who do hand out free practice material, where to practice screening formats collects what each one gives away. Every figure above is the publisher's, dated in the body and checked 7 October 2026, or it is labeled as ours.