Guide · Careers

Which resume keywords actually beat ATS filters (with examples)?

The keywords that beat ATS filters are the exact hard skills, tools, certifications and job titles pulled straight from the job description, matched at 70-80% density, in real bullets, with both the acronym and the full phrase.

Published 5 July 2026 Updated 5 July 2026 Read ~9 min Topic Careers · Resume · ATS

I get asked this weekly by friends switching jobs: "Why is my resume ghosted?" Nine times out of ten it isn't experience. It's the keyword layer. Applicant Tracking Systems parse your resume into a text blob, score it against the job description, and only the top matches reach a human. Same intent-matching logic I use in keyword research for SEO: the query is the JD, your resume is the landing page. Get the keywords wrong and nobody reads the copy. Get them right and you clear the filter. If you want to skip theory and just see your score, drop your CV and a JD into my free resume ATS checker for a number in about 20 seconds.

How ATS keyword matching actually works

Every ATS (Workday, Greenhouse, Lever, iCIMS, Zoho Recruit, Naukri RMS) does roughly the same three things. It strips your PDF or DOCX to plain text, tokenises it into words and phrases, then compares those tokens against the job description or a recruiter's search string. A "match" is not fuzzy magic. It is a token appearing in both documents, weighted by how often it appears and where. Titles and skills sections weigh more than a passing mention in a hobby line, which is why placement matters as much as presence.

Pull keywords from the JD, not from a template

The biggest mistake I see is candidates copying a "top 50 resume keywords" list from Google. That's generic. The ATS is scoring you against one specific JD. So open the exact job description, highlight every noun that names a tool, framework, certification, methodology or job title, and count how often each appears. Anything mentioned twice or bolded is required. Anything in the "must-have" or "requirements" block is non-negotiable. Everything else is nice-to-have.

Once you have that list, mirror the language exactly. If the JD says "Google Ads", do not write "AdWords". If it says "stakeholder management", do not write "worked with clients". Same skill, different tokens, and the ATS only sees tokens. When you are done, paste both documents into the resume ATS checker and it will flag every missing term along with the match percentage.

Density, exact match vs stemming, and the acronym trap

Density matters, but not the way blog posts pretend. You don't need a keyword ten times. You need it two or three times, in the right places: the title line, the skills section, and inside one achievement bullet. That signals "I have done this recently and repeatedly" without tripping the stuffing detector a human recruiter will use next. Aim for 70-80% overlap with the JD's hard-skill list. That is where interview callbacks spike in my own tests.

On matching logic: modern ATS platforms do basic stemming, so "manage / managed / managing / management" collapse to one root, and singular / plural forms are equivalent. But they do not expand acronyms. "SEO" ≠ "Search Engine Optimization". "AWS" ≠ "Amazon Web Services". "CA" ≠ "Chartered Accountant". Always write it as Acronym (Full Phrase) the first time, e.g. "SEO (Search Engine Optimization)", so both tokens land in the index. That one habit lifts scores more than any font change.

Hard skills vs soft skills: which the filter actually reads

Hard skills carry the ATS stage. Soft skills carry the interview. A filter can score "Python", "HubSpot", "GST", "Figma", "PMP" cleanly because those tokens are unambiguous and mentioned by name in the JD. It struggles with "team player", "detail-oriented" and "great communicator" because every candidate claims those and recruiters rarely search for them. Load the top of your resume with the exact hard skills from the JD (tools, certifications, methodologies, platforms, languages, frameworks, job titles) and let soft skills live in bullet points where they show up naturally through the story of what you did.

Your resume does not need to sound impressive to the ATS. It needs to contain the right tokens in the right sections. Sound impressive to the human on page two.

20 example roles and the keywords that actually matter

Below are the hard-skill keywords I would prioritise in each JD family, based on scanning hundreds of Indian and remote listings. Treat this as a starter checklist. Always cross-check against your specific JD.

  • Digital Marketing Manager: Google Ads, Meta Ads, SEO, GA4, GTM, HubSpot, CRO, ROAS, CAC, funnel.
  • SEO Specialist: on-page SEO, technical SEO, Ahrefs, Semrush, Search Console, Screaming Frog, schema, Core Web Vitals, backlinks, E-E-A-T.
  • Performance Marketer: Meta Ads Manager, Google Ads, CPL, CPA, ROAS, LTV, retargeting, lookalike audiences, attribution, A/B testing.
  • Content Writer: SEO writing, long-form, CMS (WordPress), editorial calendar, keyword research, tone of voice, briefs, AP style.
  • Software Engineer (Backend): Python, Java, Go, PostgreSQL, Redis, Docker, Kubernetes, REST, gRPC, CI/CD.
  • Frontend Engineer: React, TypeScript, Next.js, Tailwind, Redux, Jest, Cypress, accessibility (a11y), Webpack, responsive design.
  • Data Analyst: SQL, Excel, Python (pandas), Power BI, Tableau, Looker, ETL, dashboards, cohort analysis, A/B testing.
  • Data Scientist: Python, scikit-learn, TensorFlow, PyTorch, XGBoost, SQL, MLOps, feature engineering, NLP, model deployment.
  • Product Manager: roadmap, PRD, OKRs, Jira, Figma, user research, A/B testing, stakeholder management, agile, GTM.
  • UX / Product Designer: Figma, prototyping, user research, wireframes, design systems, usability testing, accessibility, Sketch.
  • Chartered Accountant (CA): GST, TDS, income tax, statutory audit, IFRS, Ind AS, Tally, Zoho Books, MIS, ROC filings.
  • HR / Talent Acquisition: sourcing, Boolean search, LinkedIn Recruiter, ATS (Greenhouse / Lever), onboarding, POSH, HRMS, employee engagement.
  • Sales / BDR: pipeline, CRM (Salesforce, HubSpot), quota, MQL, SQL, cold outreach, discovery calls, MEDDIC, closed-won.
  • Customer Success Manager: NPS, churn, QBRs, onboarding, upsell, renewal, playbooks, Gainsight, health score.
  • DevOps / SRE: AWS, GCP, Terraform, Kubernetes, Docker, Jenkins, Prometheus, Grafana, incident response, SLO/SLI.
  • Cybersecurity Analyst: SIEM, Splunk, OWASP Top 10, penetration testing, vulnerability assessment, ISO 27001, incident response, SOC.
  • Graphic Designer: Adobe Photoshop, Illustrator, InDesign, Figma, After Effects, brand identity, typography, print production.
  • Video Editor: Premiere Pro, After Effects, DaVinci Resolve, colour grading, sound design, short-form (Reels / Shorts), motion graphics.
  • Business Analyst: SQL, Excel, Power BI, requirements gathering, BRD / FRD, stakeholder management, process mapping, Jira.
  • Operations Manager: SOPs, KPIs, vendor management, P&L, supply chain, lean, six sigma, ERP (SAP / Oracle).

Notice how none of those lists start with "hardworking" or "passionate". Once you have your JD-specific list, run yours through the resume ATS checker to see which of these tokens actually landed in your parsed CV. The parser sometimes drops words hidden in text boxes or headers, which surprises people.

Where to place keywords so the parser weighs them

Placement is quiet leverage. ATS parsers weigh certain zones higher, so the same keyword in your title line beats the same keyword in a hobbies footer. My playbook: put the exact target job title from the JD at the top ("Performance Marketer") even if your last title was slightly different. Use a Skills / Core Competencies section directly under the summary with 12-18 hard skills listed as plain comma-separated tokens. Then echo the top five inside real achievement bullets with numbers ("Scaled Meta Ads spend to Rs 8L / month at Rs 22 CPL"). That mirrors the density and placement patterns Workday and Greenhouse weigh most.

Avoid the traps that quietly kill parsing: no text inside images, no keywords in the header/footer zone of Word documents, no fancy two-column templates where the parser reads left-to-right and scrambles your bullets, and no tables for the skills section. A clean single-column PDF exported from a text editor beats a designer template every time. If you want the full formatting checklist, my ATS-friendly resume guide for 2026 walks through the layout rules.

A repeatable, 20-minute keyword workflow per application

This is the exact loop I hand friends who are actively applying. It takes twenty minutes per role and it is the difference between 2% and 20% callback rates in my own tracking.

  1. Copy the JD into a plain text doc and delete the fluff (benefits, culture blurbs). What is left is the scoring rubric.
  2. Highlight every hard-skill token: tools, certifications, methodologies, platforms, job titles. Ignore soft skills for now.
  3. Tag frequency: anything mentioned 2+ times or listed under "requirements" is mandatory; the rest is bonus.
  4. Rewrite your title line to match the JD title exactly, if you honestly can.
  5. Refresh your Skills section: 12-18 comma-separated hard skills, mirroring the JD's exact phrasing (acronym + full form).
  6. Rewrite three bullets so the top five keywords show up inside quantified achievements, not skill lists.
  7. Score it: paste JD and resume into the free resume ATS checker and aim for 75%+ before submitting.
  8. Fix the gaps the tool flags, re-score, then apply.

Do that consistently for two weeks and the callback rate moves. Same "match intent, then write" logic I preach for SEO. The medium changes, the principle does not. If you want me to look over a resume and JD personally, reach out via the contact page. If you just want the automated score first, the tool is free.

Frequently asked questions

How many keywords should a resume have to pass ATS?

There is no magic count. Aim to match 70-80% of the hard skills, tools and certifications listed in the job description, not a fixed number. A senior data role JD with 12 named tools should see roughly 9-10 of them on your resume, used in real bullet points. Stuff every buzzword and a human recruiter will bin you on page one, so hit the match rate and stop there.

Do ATS systems read exact match or stemming?

Most modern ATS platforms (Workday, Greenhouse, Lever, iCIMS) handle basic stemming, so "managed" and "management" register as related. But acronyms, tools and certifications are treated as exact strings: "SEO" and "Search Engine Optimization" are two different tokens, and "AWS" does not equal "Amazon Web Services". Include both the acronym and the full phrase at least once each.

Are hard skills or soft skills more important for ATS keywords?

Hard skills win the ATS stage. Filters and recruiter searches key off tools, certifications, methodologies and job titles (Python, GST, HubSpot, PMP, Figma) because those are unambiguous. Soft skills like "team player" or "communication" are so common they rarely filter anyone in or out. Load your resume with the exact hard skills from the JD, then let soft skills carry the interview, not the shortlist.

Related guides

Score my resume free → ← Back to blog