We'd like to use Microsoft Clarity to see how visitors use this site. It records clicks, scrolling and page views. It runs only if you allow it. See our Privacy Policy.

Skip to content
Country:Global
Login

Recruiters don't read resumes. They search them.

Once you apply, you're a database record. Recruiters retrieve you with the posting's exact words. Write for the query, not just the reader.

The ApplyHustle teamEditorialUpdated 9 min read
A man in a shirt and tie working on a laptop at a desk in a small office.

The most persistent misunderstanding about resume keywords is that they exist to please a machine. They do not. They exist because, on the other side of every job posting, a recruiter is typing terms into a search box, and your resume either contains those terms or it does not.

We walked through the full pipeline in what happens in the six seconds after you hit apply. The short version is that once your file is parsed, you exist as a searchable record in a database. Resume keywords are the strings that decide whether that record is ever retrieved. Seen that way, the whole topic stops being about tricking software and starts being about matching a query.

The database behind every job posting

A live posting collects applicants faster than anyone can read them, and recruiters typically carry many roles at once. So they do what anyone with a large database does: they search it. An ATS search runs keyword queries and boolean filters against the parsed text of every candidate record, the way a lawyer searches case law rather than reading every judgment ever written.

ATS keywords are not limited to skills either. Recruiters filter on job titles, locations, employers, qualifications and years of experience, all of which are fields in your parsed record. And the searching does not stop when a role closes. When a new role opens, recruiters run queries across their whole existing database, which means the keywords in an application you sent months ago can surface you for a job that did not exist when you wrote them.

This retrieval model has a hard implication. If your record does not match the query, you were not rejected in any meaningful sense. You were never retrieved. Nobody weighed your experience and found it wanting. The search simply returned other people, and the process moved on without you.

Understanding that distinction changes where you spend effort. Polishing prose helps the reader you eventually get. Keywords determine whether you get a reader at all.

Greenhouse's help pages list the operators its candidate search accepts: AND, OR and NOT, quotation marks for an exact phrase, parentheses for grouping, and an asterisk for any ending, so analy* matches analyst, analysts and analytics. Most systems that offer Boolean search use the same grammar. A recruiter hiring a project manager for a stakeholder-heavy programme might type the string below.

  • The string: ("stakeholder management" OR "stakeholder engagement") AND (PMP OR "Project Management Professional") AND (Agile OR Scrum) NOT intern*
  • Retrieved: Led stakeholder engagement for a 14-person Scrum programme across three business units; PMP certified.
  • Not retrieved: Managed client relationships across agile teams; certified project manager.
  • Also not retrieved: the first candidate again, if their record still lists a 2016 internship, because NOT intern* drops any record containing the word.

The second candidate may have done identical work, and the string still skips them, because none of the three required groups matches their wording. The last line shows the other risk: a careless exclusion can drop experienced people for an old word on the page. You cannot control the exclusions, but you can make sure each required group has a match, in the words the posting used.

Exact-match thinking: why your synonym never gets queried

Recruiters tend to search the posting's own vocabulary, for an unglamorous reason: they wrote the posting, or sat with the hiring manager who did. The words in the job description are the words in their head when they build the query.

So if the posting says stakeholder management and your resume says client liaison, you may be describing identical work, but a search for stakeholder management will not return you. The same goes for tools and abbreviations. A recruiter searching PMP will not find Project Management Professional written out in full, and one searching Kubernetes may not think to try container orchestration.

A diagram of one search query for stakeholder management reaching two equally qualified candidate records: the record containing the exact phrase is retrieved and lifted from the pile, while the record that says client liaison is never returned at all, its dashed connector stopping short.
The synonym is not ranked lower, it is simply never returned, which is why tailoring means translating your experience into the posting's exact vocabulary.

Abbreviations deserve their own rule. You cannot know if the query will use the acronym or the full phrase, so give the record both, written once as Search Engine Optimisation (SEO) or Registered Nurse (RN). The same logic runs across markets: a Programme Manager in London is a Program Manager in Toronto, and the record that carries the posting's spelling wins the query.

The AI matching now built into some systems is more forgiving than a search box. Greenhouse's Talent Matching compares related terms on a resume with the skills a hiring team names, and SAP describes its screening as finding related skills beyond exact matches. But those skills are named from the posting, Boolean search is still exact, and you cannot know which of the two a given recruiter is using. Write in the posting's words, without inventing anything in the process, and you are covered either way.

Mining the job description for the words that matter

The job description is the search query written out in advance, which makes learning how to tailor a resume mostly a reading exercise. Pulling resume keywords from a job description takes about fifteen minutes done properly.

  • Collect the hard skills, tools and platforms named explicitly, because concrete nouns are what recruiters actually type into the search box.
  • Include the job title itself, since recruiters routinely search for people who already hold the title they are hiring for.
  • Note certifications and qualifications, and use the variant your market recognises, since a UK employer searches ACCA where a US one searches CPA.
  • Watch for repetition, because a term that appears three times in one posting is almost certainly a term the hiring manager cares about.

Do this against the posting you are actually answering, not a generic version of the role. Two postings for the same title can weight entirely different terms, and the query will be built from the one sitting in front of the recruiter.

Then separate requirements from furniture. Phrases like excellent communication skills appear in every posting and nobody searches them. A named tool, a certification, a methodology or a domain term is a plausible query. Furniture is not.

Before you finish, test your own record the way the recruiter will. Write two or three strings like the one above from the posting, one per requirement group, and search your resume text for each group by hand. Any group with no hit is a gap, and the job description has just told you the exact words that close it, if the experience is real.

Where resume keywords live: skills section vs woven into experience

Placement matters more than most advice admits. A keyword attached to an achievement reads stronger than the same word sitting alone in a skills list, for the software and the human alike. Some ranking features also consider where and how recently a skill appears, and a recruiter reading the snippet under your name sees the sentence the term sits in.

The practical pattern is to use both locations for different jobs. The skills section works as an index: a clean, scannable list that guarantees the term exists in your record. The experience section supplies the evidence: the same term inside a bullet point, doing visible work, with a result attached. A record that says Terraform in a list and then shows Terraform provisioning a production environment two lines later satisfies the query and survives the human read that follows it.

What does not work is scattering terms with no support. A skills list of forty entries with no appearances anywhere else in the record reads as padding, and screeners have seen thousands of them.

Recency cuts both ways, so use it deliberately. If your strongest match for a role sits in a job two positions back, consider whether the same skill honestly belongs in the description of your current work. If it does, say so there. If it does not, the posting may be aimed at a slightly different candidate, which is worth knowing before you spend an evening on the application.

Keyword stuffing: the trick that stopped working

The old trick was white text: paste the entire job description into the document in a font the eye cannot see, and let the machine count the matches. It fails today on mechanism, not morality. Recruiters screen the parsed record, and parsing strips all formatting, including colour. Your invisible paragraph surfaces as a block of visible gibberish at the bottom of your profile.

Repeating a keyword nine times fares no better. The parsed record makes repetition obvious, and the human in the loop, the person examined in who actually rejected your resume, discards records that look gamed. Stuffing does not beat the filter. It converts a possible keep into a certain discard, and the modern variants, keyword blocks in tiny fonts or crammed into margins, fail the same way, because anything the parser can read, the recruiter can see.

Checking coverage with a match score

The honest version of keyword matching is coverage checking: put your resume next to the posting and see which terms are present, which are missing, and which are present but poorly placed. ApplyHustle's free ATS check does that when you give it the job description, listing the keywords and requirements the posting asks for, which your resume shows, and where. In the Jobs feed, AI Match Score does the same job one level up, scoring your profile against each posting and listing the skills it asks for that you have not shown.

The point is not to chase a perfect number. Gaps come in two kinds. Vocabulary gaps are fixable in minutes, because you already do the work and named it differently. Real gaps, skills you do not have, are information of a different sort, the kind that feeds the decision covered in should you even apply. Run the check before you apply rather than after the silence, because coverage is only useful while you can still act on it.

Keywords are not a trick to be played on software. They are the shared language between a posting and a record, and the candidates who get retrieved are simply the ones who bothered to speak it.

Questions about resume keywords

Do recruiters really use Boolean search in an ATS? Many systems offer it, Greenhouse among them, and recruiters use it most when a role draws hundreds of applicants or when they reopen an old talent pool. Others filter on parsed fields or lean on a built-in AI ranking, and the same posting vocabulary serves all three.

Does an ATS understand synonyms? A Boolean search does not. AI matching features compare related terms, but they match against skills named from the posting, so the posting's wording is still the safest choice.

How many times should a keyword appear? Once in the skills section and once in a bullet that shows you using it is enough for retrieval. More than that only helps if each mention is a different piece of evidence.

Share this article

Keep reading