A Warehouse Supervisor Tested His Nine-Second Rejection

Marcus kept a spreadsheet with one row for every application. The columns covered the employer type, job title, pay range, résumé version, response and days waiting. He added another column after the rejection that came nine seconds after he pressed submit.

He called it “minutes to rejection,” though the first entry was a fraction of one.

Marcus is a composite of warehouse workers who described applying through automated hiring systems. He had supervised 26 people on a distribution shift, trained new hires, handled inventory discrepancies and filled in when equipment operators called out. After his employer reduced hours, he began looking for another supervisor job within a 30-mile drive.

Over six weeks, he applied to 40 openings. Four rejection emails arrived in under a minute. The fastest came after nine seconds, before he had closed the application tab.

The notice said the company was moving ahead with other applicants. It did not say whether a person had opened his résumé, whether a required answer had removed him, or whether a score had placed him below a cutoff. That missing distinction became the point of the spreadsheet.

The nine-second row

Marcus opened the job posting again. The work resembled what he had done for five years: assigning labor, checking shipments and documenting safety issues. The posting used the title “warehouse supervisor.” His résumé called his latest role “operations lead,” the title printed on his pay statements.

One application question had asked whether he held a college degree. He had selected no. The posting described a degree as preferred, not required.

Either detail could have mattered. Neither detail may have mattered. The email’s speed offered no answer.

An automated hiring system can reject an application quickly when an employer has set a firm rule around a response, such as work authorization, location or a required credential. It can also assign a low match score after comparing the application with words, titles and qualifications in the posting. Some systems sort candidates into groups rather than rejecting them outright, leaving an employer to decide which group receives attention.

The elapsed time between submission and email does not reveal which process occurred. A system may calculate a result as soon as the application arrives, while the rejection message may be released immediately or held for later. A nine-second email shows that automation was possible. It does not prove that the full decision took nine seconds, and it does not reveal whether a recruiter had set the rule weeks earlier.

Marcus entered “9 seconds” in the spreadsheet anyway. It was the only evidence he had.

What a screening score measures

A screening score is usually not a general rating of a worker. It reflects the information available to the tool and the instructions attached to one opening.

Those instructions may include required answers, years of experience or a set of skills drawn from the job description. A system might give weight to recent use of a phrase, recognize related terms, or treat a specific job title as a closer match than another title covering similar work. If an employer adds an assessment, the result may become another input.

The number can look more complete than it is. A score of 62, for example, does not mean a person has 62 percent of the ability needed for the job. Depending on the tool, it may mean that the application matched part of a configured profile, landed in a certain rank among applicants, or failed to collect points in fields the employer chose to emphasize.

A résumé supplies limited evidence. Marcus’s document showed that he had supervised a shift and handled inventory, but it did not contain the phrase “cycle count,” which appeared in several postings and described work he performed each month. His official title did not contain “warehouse.” A parser could also place text in the wrong field when it encountered columns, tables or unusual headings.

None of this requires a system to understand whether Marcus could settle a conflict between two workers, notice that a loading pattern was unsafe, or keep orders moving after an equipment failure. Unless the application captured those facts in a form the system could use, they were absent from the score.

Employers can use screening tools to manage hundreds of applications, enforce stated requirements and bring likely matches closer to the top. The same structure can hide consequential choices. A preferred credential may operate like a requirement. A title associated with one employer may receive less weight than a common title.

Past hiring patterns can shape a model or its evaluation targets, carrying old differences into a new ranking.

Federal employment protections still apply when software participates in hiring. Automated tools can create discrimination concerns when their criteria disproportionately screen out protected groups, including people with disabilities. Disclosure rules differ by location, and many applicants receive little more than a standard rejection notice.

The résumé test

After 12 applications, Marcus stopped sending the same résumé each time. He saved two versions and marked each one in the spreadsheet.

The first kept his official title and the two-column layout he had used for years. The second added a plain description beside that title: warehouse supervisor. It moved his skills into ordinary lines of text and replaced broad phrases with work he had done, including cycle counts and forklift training.

He did not add credentials or change his employment dates. The facts stayed put. Their labels changed.

This was not a controlled experiment. The jobs came from different employers, attracted different applicant pools and may have used different settings. Even openings with the same title could place different weight on weekend availability or the number of direct reports. Marcus could observe outcomes, but he could not isolate a single cause.

The pattern still changed. Among his first 12 applications, four produced rejection emails in under a minute and none led to an interview. Across the next 28, using the revised résumé for 21 of them, he received five recruiter screens and two interviews. One interview came from an application that remained unanswered for 11 days before a person contacted him.

The spreadsheet did not show that the revised résumé beat an algorithm. It showed that clearer overlap between his record and the posting coincided with more human contact, while leaving open the effects of timing, competition and employer settings.

One row made the uncertainty plain. Marcus used the revised résumé for a supervisor opening that matched his experience, then received another fast rejection. On rereading the posting, he noticed a requirement for overnight availability. He had marked that he could not work overnight because he shared care of his daughter.

That rejection may have measured availability with perfect accuracy. It said nothing about his warehouse work.

The human choice inside the tool

Automation can make a rejection feel ownerless, but people decide what the system receives and what happens to its output. An employer chooses which questions are mandatory, how a job description is written and whether recruiters review candidates below a score.

Those choices are often invisible to the applicant. Marcus could see a submission confirmation, a status change and an email. He could not see the threshold or learn whether “degree preferred” had become a weighted preference, a hard filter or nothing at all.

A recruiter may also use a score as one signal rather than a final decision. Another employer may treat the ranked list as the working applicant pool and never open the rest. The same tool can play different roles because its settings, staffing and local practices differ.

That is why the phrase “rejected by an algorithm” can be both understandable and incomplete. Software may execute the decision, but the decisive condition could come from an employer’s rule. In other cases, the employer may rely on a model whose ranking cannot be explained to the applicant in ordinary language.

Marcus never learned which applied to the nine-second row.

He kept tracking applications after accepting another warehouse job eight weeks later. The offer came from one of the revised résumé submissions, following two conversations and a walk through the work area. The pay was $2.25 more an hour than his reduced rate, though the commute added 14 miles each day.

In the spreadsheet, he marked the outcome “accepted.” He left the nine-second rejection unchanged.

Questions people ask

Can an employer reject an application automatically?

Many employers configure hiring systems to remove applicants who do not meet selected conditions or to rank applications before a recruiter reviews them. The rejection may be sent without a person opening the résumé, although the employer generally chose the questions, requirements or thresholds that produced the result.

Does a fast rejection mean the résumé was never read?

It may mean no recruiter reviewed it, but the email’s arrival time cannot establish that by itself. A tool may parse and score a résumé immediately, apply a knockout answer, or release a decision that was already determined by the employer’s settings.

Is a screening score a measure of job performance?

Usually it is a measure of match under a particular system. It may reflect résumé language, application answers, assessments or employer-selected criteria, while leaving out parts of the work that were never captured. Marcus’s score could recognize “warehouse supervisor” more readily than his official title, even though his experience had not changed.

Do employers have to explain an automated rejection?

Disclosure requirements vary by location and by how the tool is used. Many applicants receive no score, cutoff or specific reason, even where employment protections still cover automated decisions. Marcus’s only record was the standard notice and the “9 seconds” entry in his spreadsheet.