Greenhouse Resume Checker: Keywords, Parsing & Rejection

When you hit submit on a Greenhouse job board, two things happen. Your resume file is stored exactly as you sent it, and a parser extracts your contact details, work history, and education into a searchable candidate profile. That profile lands in a recruiter's queue. The parser has documented limits: per Greenhouse's own support docs, it can't process files over 2.5MB and struggles with tables, columns, graphics, word art, and contact information placed in headers or footers. When parsing fails, your searchable profile is incomplete — even though your original file looks fine.

Greenhouse's current product has more than one review mode. Talent Filtering searches application text using exact Required and Preferred terms. The optional Talent Matching feature compares resumes and application answers with employer-defined, weighted calibration criteria and groups candidates into Strong, Good, Partial, Limited, or Needs manual review. Greenhouse describes this as assistive AI: it does not automatically advance or reject candidates, and recruiters can review or override its output.

In Talent Filtering, recruiters can search full resume and note text for job titles, skills, locations, and other terms. Greenhouse documents literal matching: the keyword in the search must exactly match the application. Required terms behave like AND; Preferred terms behave like OR. Talent Matching is different: it can highlight exact and similar terms and uses the hiring team's weighted skills, experience, job titles, and optional industry criteria. Whether either workflow is enabled depends on the employer's tier and configuration.

One genuine automation does exist: application rules, often called knockout questions. An employer can configure a job so that a specific answer to a custom question — Yes/No, single-select, or multi-select — triggers automatic rejection. "Are you authorized to work in this country?" is the classic example. This keys off your answers, never your resume content, and each employer configures it per job. Whether a given role uses knockout rules varies and isn't visible to you as a candidate.

Greenhouse describes the reality plainly: your application goes "straight into a queue on a recruiter's dashboard," and recruiters can view your submitted resume alongside your form answers. On high-volume roles, Talent Filtering can narrow the pile by exact job-title, skill, location, or other keyword matches plus location, referral, scorecard, education, and custom filters. If the relevant wording is absent or the document parsed incompletely, your application may not appear in that particular filtered view even though the original file remains attached.

The practical workflow follows directly from Greenhouse's documentation. Read the public job post, identify the exact titles, tools, certifications, methods, and domain terms you genuinely possess, and use those phrases naturally in your summary, skills, or experience. That supports literal Talent Filtering and gives criteria-based Talent Matching better evidence. Do not repeat a phrase for density: neither source documents a public candidate-facing keyword-frequency pass mark.

The real rejection mechanism is a human moving fast, not an algorithm. Eye-tracking research from Ladders puts the average first resume screen at 7.4 seconds — so your resume has to survive a skim, then a scorecard. And silence usually isn't rejection-by-robot either: candidates filtered out on a busy role often simply never get reviewed, and the closure email comes late or not at all, depending on the employer.