Keyword matching is the single biggest lever in resume screening — in our Merit Score, skills & keywords carry a 25% weight, more than any other axis. But most keyword advice is stuck in 2015: copy words from the job description, paste them into your resume, done. Modern matching is smarter than that, and so is the failure mode.
25%
of the score is skills & keywords — the largest axis
10%
more rides on semantic coverage of JD requirements
5%
on matching the job title itself
Not all keywords are weighted equally
A job description is not a flat list of words. Matching engines (ours included) weight terms by where and how they appear:
Job description — extracted & weighted
Under “Requirements”
Under “Nice to have”
Company blurb & perks
- Requirements beat perks. A skill under must have or requirements counts far more than one buried in nice to have or the company blurb.
- Repetition is signal. A tool mentioned four times across the JD is core to the role; a tool mentioned once in passing is set dressing.
- The title is a keyword. If the posting says Senior Data Analyst and your last title says Data Analyst, that gap is measured too — title fit is its own scored axis.
So before you touch your resume, rank the JD's terms: hard requirements first, repeated tools second, everything else last. Cover the top of that list and stop.
A repeatable way to pull keywords from a posting
Fifteen minutes, no tools required:
- 1
Copy the posting into a blank document
Strip the company boilerplate, benefits and EEO statement — none of it is scored.
- 2
Mark every concrete noun
Tools, languages, platforms, certifications, methodologies, domains. Ignore adjectives entirely.
- 3
Tally repeats and note the section
A term in 'Requirements' that appears three times is your top priority. One mention in 'Nice to have' is your last.
- 4
Split into have / partially have / don't have
The first group needs surfacing, the second needs evidence, the third you leave alone.
That third column is the one people get wrong. The correct action for a skill you don't have is to leave it off — not to claim it, and not to abandon the application. Most postings list more than any real candidate has.
Hard skills match. Soft skills mostly don't.
"Team player", "excellent communicator", "detail-oriented" and "self-starter" are nearly worthless as keywords. They appear in every posting and every resume, so they carry almost no discriminating signal, and no matching engine can verify them.
That doesn't mean communication doesn't matter — it means the resume proves it differently. "Presented quarterly roadmap to a 40-person org" demonstrates communication; the phrase "excellent communicator" asserts it. One is evidence, the other is filler that costs you a line.
Synonyms and implied skills usually count
Good matching engines expand keywords through a skills taxonomy. Ours knows that K8s is Kubernetes, that React implies JavaScript, and that Postgres satisfies a generic SQL requirement. That means you don't need every surface form — but you do need the specific term when the JD is specific. If the posting says Snowflake, a generic data warehousing bullet is a weaker match than the literal word.
The same pattern applies beyond tools: Search Engine Optimization (SEO), Certified Public Accountant (CPA), Registered Nurse (RN), Objectives and Key Results (OKRs). Write it out once where it naturally belongs, then use the short form everywhere else.
What to do when your job title doesn't match
Internal titles are a real and common problem. "Customer Happiness Ninja", "Associate II", "Member of Technical Staff" and "Analyst III" all describe work the market calls something else, and title fit is scored on its own.
The honest fix is a parenthetical, not a rewrite:
Unmatched internal title
“Customer Happiness Ninja · Acme Corp · 2022–2025”
Scores near zero on title fit and tells a recruiter nothing.
Same title, market-legible
“Customer Happiness Ninja (Customer Success Manager) · Acme Corp · 2022–2025”
Keeps the real title on the record while matching what the market searches for.
You are not allowed to change what your title was — that's a verifiable fact and background checks do catch it. You are entirely allowed to translate it.
Where keywords live changes what they're worth
A keyword in a skills list proves you claim the skill. A keyword inside an experience bullet proves you used it — attached to an employer, a date range, and ideally a number. Semantic coverage in our engine is section-weighted for exactly this reason. The strong pattern:
- 1
List it in your skills section
Exact matchers find it instantly — this is the floor, not the ceiling.
- 2
Back it with an experience bullet
The same term inside a dated role, showing the skill in use, with an outcome attached.
- 3
Pull the top 2–3 into your summary
The JD's most critical requirements should be visible in the first three lines a recruiter reads.
Keyword stuffing fails on contact with a human
It also does nothing for the axes that measure evidence: years-per-stack, quantified outcomes, seniority trajectory. Add only keywords you can defend without lying.
A useful sanity check before you submit: read your skills section out loud as if a hiring manager just asked "tell me about your experience with each of these". Anything you'd have to hedge on shouldn't be there.
See your actual match rate
The free MeritSlate checker extracts the weighted keywords from a real job description, shows which ones your resume hits — including synonym and implication matches — and lists the gaps in priority order. Fix the top of the gap list first; it is where the points are.
Frequently asked questions
- How many keywords should I put in my resume?
- Cover the job description's hard requirements and its repeated tools — usually 8–15 terms — and stop. Beyond that you're adding words you can't defend in an interview. Coverage of the top of the requirement list beats raw keyword count in every modern matching engine.
- Does keyword stuffing still work?
- No. Parsers normalize white text, hidden layers and keyword walls into the same plain text a recruiter reads, so stuffing survives the machine and destroys your credibility with the human. It also adds nothing to the axes that measure evidence, like quantified outcomes and years per stack.
- Should I write acronyms or spell skills out?
- Write both, once: "Kubernetes (K8s)". Taxonomy-based matchers understand either form, but older exact-match systems only find the literal string in the job description. One naturally written line covers both worlds without looking like padding.
Written by
MeritSlate Team
The team building MeritSlate's deterministic scoring engine — nine evidence-anchored axes, a nine-vendor ATS parse simulation, and every point backed by receipts.