The assessment is the employer brand.

Hiring conceals an audience problem. For every candidate an organization hires, the selection process was experienced in full by everyone it turned away, and that turned-away majority is the largest group of people the company will ever address one at a time. Few of them will ever see the office, meet the manager, or use the product. What they saw, closely and under pressure, is the assessment — which makes candidate experience less a courtesy extended to applicants than the widest broadcast the employer runs.

Most applicants never encounter the company at higher resolution than in the assessment itself. A careers page is read at a distance, and an interview is a conversation reserved for a small pool; the assessment is an hour or two spent inside a process the organization designed, under rules the organization chose, observed in ways the organization decided. Every one of those choices is legible from the inside. Candidates cannot see the validity study behind an instrument, but they can see whether the tasks resemble the job, whether the rules applied to them applied to everyone, and whether anyone told them what would happen next.

Industrial-organizational psychology has studied that inside view for three decades under the name applicant reactions, and it has mapped what the candidate's verdict runs on with unusual clarity. This article gives that literature its full hearing: the rules candidates judge by (Gilliland, 1993), where their judgments travel once formed (Hausknecht, Day, & Thomas, 2004), the inexpensive lever most programs never pull (Truxillo, Bodner, Bertolino, Bauer, & McCarthy, 2009), and the places where the evidence is thinner than the topic's advocates admit (McCarthy et al., 2017). It closes on the point the series has been building toward: everything this literature asks employers to provide, the preceding articles already recommend on validity, fairness, and privacy grounds alone.

The rules candidates judge by were catalogued three decades ago

The framework the field still works from appeared before online assessment existed. Gilliland (1993), writing in the Academy of Management Review, took organizational justice theory, the branch of organizational psychology that asks why procedures feel fair or unfair regardless of who wins, and applied it to selection systems. The move separated two judgments that hiring conversations tend to blur: whether the outcome was fair (I deserved the offer) and whether the process was fair (the way they decided was defensible). Candidates render both, and the second is the one the employer fully controls.

Gilliland catalogued the procedural rules that drive the second judgment: job-relatedness, consistency of administration, opportunity to perform, explanation and feedback, selection information (being told in advance how the process will work), reconsideration opportunity, truthfulness of communication, interpersonal effectiveness, two-way communication, and the propriety of questions. The names are academic; the content is not. In the candidate's language, the rules ask whether the process looks like the job, whether it ran the same way for everyone, whether there was a real chance to show what one can do, whether anyone said what was happening and why, whether the treatment was respectful, and whether questions went answered rather than into a void. Figure 1 lays the main rule families out in exactly those terms.

Two properties of the catalogue matter for anyone running an assessment program. First, the rules are observable without expertise. A candidate cannot audit an item bank, but job-relatedness, consistency, and respect are exactly the properties a stressed non-specialist can read from a testing session, and they are the only data the candidate has. The verdict gets formed either way; the only question is whether anything was designed for it. Second, Gilliland's central proposition is that satisfying the rules shapes the perceived fairness of the process and of the outcome together. Most applicants receive an adverse outcome, and the fairness of the road to it is a large part of how the destination is received. An organization that cannot change its rejection rate can still change what its rejections are made of.

It is worth being precise about what the framework is. Gilliland's paper is a theory with a catalogue inside it, not a measurement study, and its numbers-free character is why this article prints none. What came after it was a sustained empirical program testing whether the catalogued perceptions actually track anything. The answer is the next section.

The verdict leaves the room with the candidate

Hausknecht, Day, and Thomas (2004) aggregated that empirical work in Personnel Psychology, in the meta-analysis that mapped where applicant reactions go once formed. The pattern is consistent. How candidates perceive selection procedures relates to how attractive the organization looks to them, to their stated intentions toward it, and to their perceptions of themselves. Not one of those destinations stays inside the hiring funnel.

The intentions cluster deserves unpacking, because each intention sits one step upstream of a quantity the business already tracks. Intention to accept bears on offer conversion: the finalists an organization most wants are the ones most likely to hold competing offers, and the process is part of what they weigh. Intention to recommend bears on word of mouth: candidates describe selection processes to friends, colleagues, and public review platforms, and the perceptions this literature measures are the raw material of those accounts. Intention to reapply bears on the future pipeline: this quarter's silver medalist is next year's stronger applicant only if the first experience left the door feeling open. What the evidence establishes is the intention, not the behavior behind it, and Figure 2 draws the map with its rightmost region deliberately faint for exactly that reason.

The fourth solid path, reception of the outcome, is Gilliland's proposition observed in the wild: the same rejection lands differently depending on the process that produced it, which makes process design a form of rejection design whether or not anyone names it that. The self-perceptions path is the one buyers overlook, and it is the least comfortable. A selection process tells candidates something about themselves, and a process experienced as arbitrary or opaque can send a person away thinking less of their own abilities on the strength of a measurement that measured badly. That cost never appears in a funnel metric. It is also the clearest reason experience is an ethical property of assessment rather than a marketing one: the instrument is writing on people, whichever way the decision goes.

Hausknecht and colleagues also found signal in which procedures fare well. Procedures perceived as more job-related, work samples and interviews among them, were received more favorably than procedures perceived as less job-related. Candidates, in other words, reward visible relevance. The favorability ranking is not a ranking by pleasantness or production values; it tracks whether the person can see the job inside the exercise. That single result carries most of this article's practical weight, because perceived job-relatedness is a property employers can design for directly, and because the same property anchors the validity case examined across this series, starting with the flagship review of what predicts job performance.

The cheapest lever is an explanation

Within the rule catalogue, one family is so inexpensive to satisfy that its neglect needs explaining: telling candidates what is happening and why. Truxillo, Bodner, Bertolino, Bauer, and McCarthy (2009) reviewed the accumulated studies in a meta-analytic review in the International Journal of Selection and Assessment, and the finding is as clean as this literature gets: providing applicants with explanations, about the procedure, about its job-relatedness, or about the decision, improved fairness perceptions and related reactions across studies. The lever works at both ends of the process. Before the sitting, an explanation is notice: what will be measured, why it relates to the role, what will be observed during the session. After it, an explanation is closure: what happens next, on what timeline, and what the outcome was. The lever sits where the field always said the perceptions form: applicant reactions ride on how the process is explained and administered, not only on what the procedure contains (Ryan & Ployhart, 2000).

What makes the result strategically interesting is the asymmetry between cost and visibility. The properties organizations spend most on, item quality, calibration, scoring models, are precisely the ones candidates cannot see; the explanation, which candidates experience directly, costs a few carefully written paragraphs and an operational commitment to send them. Yet the modal candidate experience remains an unexplained battery followed by silence, because explanation sits in nobody's budget line and its absence shows up in nobody's dashboard. The reactions literature prices that silence: candidates fill the vacuum with the least charitable reading, and the fairness perception decays accordingly.

Explanation also compounds with the outcome. Because most explanations reach people who are about to be rejected, they are cheap precisely where the audience is largest, and the perceived fairness they buy is applied against the process's hardest moment, the adverse decision. An organization that explains its procedure up front and its outcome at the end has addressed the rejected majority twice, respectfully, in its own voice, for the price of two well-written documents. Very few line items in a talent budget reach that audience at all.

What the reactions evidence does not claim

The measured version of this case needs stating, because the topic attracts overstatement, and the field has audited itself against that temptation twice. Ryan and Ployhart (2000), in the earlier of its two critical self-reviews, insisted that the claims stay sorted: evidence about applicant perceptions is not yet evidence about downstream behavior, and research that blurs the two flatters itself.

McCarthy and colleagues (2017), reviewing the field in the Journal of Applied Psychology under the pointed questions "So what?," "What's new?," and "Where to next?," conclude that reactions matter most clearly for perceptions, intentions, and candidate-side outcomes; effects on hard organizational outcomes are more modest and context-dependent. Hausknecht, Day, and Thomas (2004) drew the same boundary around their own results: relations with behavioral outcomes are less established than relations with perceptions and intentions. The dashed region of Figure 2 is that boundary drawn as geometry, and any business case built on this literature should be built on the solid region. The boundary is no indictment: following rejected candidates into their subsequent decisions and careers is genuinely difficult research, and the field has been more candid about the gap than most of the industry quoting it.

The second limit is harder and belongs in every conversation about candidate experience: no amount of experience polish rescues an invalid instrument. A delightful process that measures nothing is theater with good manners, and the candidates it delights are being sorted by noise. Validity is the load-bearing wall, and the evidence on what actually predicts performance is the subject of the series flagship; everything in this article assumes that wall is standing. Experience is what the reactions literature says it is: the property that determines how the measurement is received, not whether it measures.

Everything candidates ask for is already on the build list

Read Gilliland's catalogue next to this series' recommendations and a pattern emerges that is too systematic to be coincidence. Each demand the justice literature says candidates place on a selection process is met by a practice the preceding articles already recommend, and recommend for reasons that have nothing to do with how candidates feel. The convergence is the strategic finding, and Figure 3 draws it in full, once, as the closing argument of the series.

Walk the rows. Job-relatedness is the experience-side name for validity-anchored content: the job-related methods candidates receive best sit at or near the top of the predictive rankings in the flagship review. Consistency of administration is the experience-side name for standardization, which the structured-hiring article recommends because holding the process constant across candidates is how bias is squeezed out of it. Explanation and feedback are this article's own territory: notice before the sitting and closure after it. Respectful observation is proportionate monitoring, which the proctoring article argues for because announced, stakes-matched observation is what deters. Care with personal data is governed data handling, which the candidate-data-privacy article requires as a matter of law and stewardship before anyone asks about perceptions. And a fair hearing when software raises suspicion is the human review of integrity flags that the article on false positives demands on statistical grounds, with the added reminder that how an organization treats the candidates its software suspects is watched by every candidate it does not.

The convergence means an organization never has to choose between the process that selects well and the process that is received well; the good process and the well-received process are the same buildout. It also means candidate experience needs no separate budget, no parallel initiative, and no trade-off analysis against rigor. The rigor, done properly and then explained, is the experience. What the reactions literature adds is a reason to finish the job: the parts of the buildout that face the candidate, the notice, the explanation, the delivered outcome, are the parts many programs leave unbuilt because no psychometric argument forced them.

Candidate experience is designed, not messaged

The convergence stays abstract until it is translated into what a candidate actually encounters, so here is the assessment candidate experience the evidence describes, stated as build decisions. There is nothing sentimental in the list: each item is a justice rule translated into product behavior, and each earns its place on the evidence above rather than on brand aspiration.

• Content that visibly resembles the job. The place to start, because it satisfies the property candidates visibly reward (Hausknecht, Day, & Thomas, 2004) with the same content decision that carries the validity case. A candidate who can see the role inside the exercise needs no persuading that the exercise is fair.
• A sitting sized to the decision. Length is an experience variable candidates read as respect for their time, and the case for measuring precisely in fewer items is made in the adaptive-testing article; it needs one clause here, not a reprise.
• Plain-language notice before the sitting. What is measured, why it relates to this role, what is monitored during the session, and what happens to the data afterward. This is Truxillo and colleagues' procedural explanation, delivered where it moves perceptions: before the candidate commits an evening to the process.
• An explanation at the end rather than silence. The decision, its timing, and what follows. The rejected majority is the main recipient, which is the point: closure is the one deliverable every candidate receives.
• Outcomes delivered, not ghosted. "Answered, not ghosted" is a justice rule wearing plain clothes, and it is the cheapest row of Figure 1 to satisfy. A process that ends with a delivered outcome has kept its last promise, and the last promise is the one candidates remember.

Two details separate programs that mean this from programs that gesture at it. The first is where the writing effort goes: the candidate-facing notices deserve the same drafting care as the hiring manager's report, because each document is the product for its reader, and the candidate's document has the larger circulation. What a well-built report contains is visible in the sample report; the discipline is to hold the pre-sitting notice to the same standard. The second is that every item above is an assessment-platform behavior before it is a communications behavior, which is why it belongs in platform evaluation criteria rather than in an employer-branding review. A vendor that cannot show you the candidate's screens is asking you to buy half the process sight unseen.

Treat the rejected majority as the audience it is

The implications amount to a change of accounting. Assessment programs measure themselves on the hired: quality of hire, time to fill, funnel conversion. The reactions literature adds the rest of the cohort to the ledger, and Figure 4 draws the proportions no dashboard shows. For each role, the thin slice at the end of the bar received an offer and a relationship; everyone else received only the process, and every one of them left with a formed judgment of the company that ran it.

Accounting for that audience starts with instrumentation. A program that tracks validity evidence should track experience evidence beside it, and the observable kind, not the aspirational kind: where in the sitting candidates stall or quit, what questions they send during and after the process, what themes recur in complaints and in public commentary. No survey vendor is required, only a willingness to read the exhaust the program already produces: the platform logs where sittings stall, and the inbox archives what candidates could not find out on their own. Perceptions of hiring process fairness move with what those traces reveal, and a recurring complaint theme is a justice rule failing in production. Programs that want to ask candidates directly can do that with published rigor too: Bauer and colleagues (2001) operationalized Gilliland's justice rules into a validated measurement scale, the Selection Procedural Justice Scale, which means candidate experience is a measurable property of a selection system rather than a sentiment about one.

It ends with a discipline about what "good experience" means. The market sells engagement, and an assessment can be made engaging the way a game is engaging, while measuring less with every added flourish. The literature ranks the properties: candidates respond to perceived job-relatedness, and no finding in this article licenses trading relevance for delight. When a program must choose where the next unit of effort goes, the reactions evidence and the validity evidence point at the same place, which is the whole argument of Figure 3. Build the process the evidence recommends, then show the candidate that this is what they are inside: name what is measured, run it the same for everyone, watch in proportion, explain, and answer. The rejected majority will never read the technical manual. The process is the only publication they receive, and it is the one the organization controls to the last word.