How to Pick Keywords From a Job Description (and Which Ones Matter)

To pick keywords from a job description, read it in five passes:

1. Split the posting into requirements, responsibilities and everything else (company blurb, benefits). Keywords come almost entirely from the first two.
2. Highlight the hard terms: skills, tools, certifications, domains, and the exact job title.
3. Mark what repeats or comes first. A term named three times, or in the title, is load-bearing.
4. Separate must-have from nice-to-have. Words like 'bonus' or 'a plus' mark the list you can skip.
5. Keep the 10–15 you can prove, and put each one inside a bullet with a result, not just in a skills list.

Most advice stops at 'use words from the job description'. That's true and almost useless. A posting contains hundreds of words. Only a small set of them decide whether your resume reads as a match, and stuffing in the rest makes it worse. The rest of this guide is about telling the two apart.

What a keyword actually is here

A keyword is a term a human screener or a search filter would use to decide whether you can do this job. That is narrower than "a word in the posting". "Collaborate", "fast-paced" and "passionate" appear in almost every listing ever written, which is exactly why they carry no signal — a term that fails to distinguish any candidate from any other cannot help you.

The terms that do carry signal fall into five groups, and they are not equally important.

Which job description keywords matter most

The ranking matters more than the list. When your resume has room for eight new terms and the posting suggests forty, taking the top three rows is close to optimal and taking the bottom row is close to worthless.

For ready-made lists of what these look like in practice, see resume keywords examples for 10 roles. Engineers have their own breakdown by specialization in the software engineer resume keywords guide.

How to read a posting for keywords

Work through the posting once with a specific question in mind for each pass.

1. Repetition. Anything named three or more times across the summary, responsibilities and requirements is load-bearing. Writers repeat what they care about without meaning to.
2. Position. Requirements listed first were usually written first, and were usually the reason the role opened. A tool named in the job title itself is the single strongest term in the document.
3. Modality. Separate "required" and "must have" from "nice to have", "bonus" and "a plus". The second list is where candidates waste effort.
4. Specificity. Prefer the concrete over the generic: "Snowflake" over "data warehousing", "SOC 2" over "compliance experience". Specific terms are the ones a recruiter actually searches for.

If the same concept appears in two vocabularies — "experimentation" in the summary and "A/B testing" in the requirements — use the one from the requirements section. That is the wording most likely to have been copied into a search filter.

A worked example: one posting, marked up

Here's a shortened product manager posting, typical of the PM roles I read as a product manager myself, and what to take from it.

Senior Product Manager, Payments

Own our checkout and payments roadmap. Run discovery with customers,
define requirements with engineering and design, and launch experiments
to improve checkout conversion.

Requirements
- 5+ years of product management in B2B SaaS or fintech
- Experience with payments, checkout or billing products
- Strong SQL; experience running A/B tests
- Track record of shipping with cross-functional teams

Nice to have
- PCI DSS knowledge; Stripe or Adyen experience

Eight keepers out of roughly ninety words, and two of them (payments and checkout) carry most of the weight. A resume for this role that never says 'payments' in a bullet will read as a near-miss to the recruiter, however strong it is otherwise.

Compare three to five postings for the same role

One posting tells you what one hiring manager wrote. Three to five postings for the same title tell you what the market expects. Paste them into one document and count: a term that shows up in four of five postings belongs in your resume's core, while a term in only one is that employer's preference and goes in only when you apply there.

That split is the practical answer to tailoring without rewriting from scratch. Your base resume carries the market's keywords; each application adds the two or three that posting alone cares about.

Where a keyword has to appear to count

Placement changes what a term is worth. The same word can be strong evidence or an unsupported assertion depending on where you put it.

A rough hierarchy of where a term is worth most: in a bullet with an outcome attached, then in a job title, then in a project description, then in a skills list, then nowhere. Terms that appear only in a skills list are the ones interviewers most often find you cannot actually speak to.

The stuffing trap

Two things happen when you push keyword density past what your experience supports. The obvious one is the interview: you get asked about Kafka because your resume says Kafka, and the conversation ends badly. The less obvious one is that a resume optimized for term coverage reads worse to the human who makes the decision — the bullets flatten into lists, the outcomes disappear, and the document stops making an argument.

White text, hidden divs and keyword blocks in a footer are worse still. They are trivially detected, and a hiring team that finds one usually stops reading. Whatever this trick was worth years ago, it is now a downside risk with no upside.

The honest version of keyword optimization is narrower and duller than the advice usually implies: find the terms you genuinely have evidence for, make sure that evidence is visible, and leave the rest alone. Gaps you cannot close truthfully are information — they tell you which roles to target and what to learn next. For how to spot them, see resume vs job description.

Doing this without reading every posting by hand

The mechanical part — extracting terms from a posting and checking which appear in your resume — is what a keyword scanner is for. The free ATS resume checker does it against whatever posting you paste: matched terms, missing terms, and which of the missing ones are load-bearing rather than filler. No account needed.

What no tool decides for you is which gaps are worth closing. A missing term you have real experience of is a writing problem and takes five minutes. A missing term you have never touched is a career-planning question, and no amount of rewording answers it. That is also why the headline percentage matters less than the list beneath it — see what a good ATS score actually means.

If you would rather have the whole document assessed than matched against a single posting, the free resume gap analysis grades it end to end. And once your evidence is captured properly in an Achievement Bank, covering a new posting's vocabulary stops being a rewrite and becomes a selection problem.