ATS & filters · 2026-08-14 · 15 min read

ATS keywords on a CV (without stuffing)

How to pull keywords from the job ad and place them in summary, achievements, and skills without stuffing. A 20-minute method and mistakes that backfire. ClarCV.

Keywords are the bridge between the posting language and your experience. An ATS uses them to score or filter; a recruiter uses them to confirm fit in seconds. The most common mistake is not using too few keywords — it is using them badly: footer stuffing, empty repetition, or invented terms that will not survive the interview.

This guide gives a practical method to extract keywords, prioritize them, place them with context, and check whether your CV already covers them. You will speak the role language without lying and without turning the document into buzzword soup. ClarCV helps spot vagueness and regenerate a clean version where skills appear as real text.

File structure (not only keywords) is in how to make a CV for ATS filters.

What a useful keyword is (and is not)

A useful keyword describes real work: tools (Salesforce, Excel, Figma), methods (Scrum, OKRs), domains (B2B SaaS, retail), responsibilities (onboarding, forecasting), and required certifications. A useless keyword is a generic adjective (proactive), a slogan (results-oriented), or a trend with no evidence (synergy, AI when it does not apply).

Prioritize terms in must-have requirements and in the first third of the posting. Nice-to-haves matter less. If the ad repeats customer success and churn, those are anchors; young culture is not.

• Named tools and stacks.
• Responsibility verbs for the role (not only soft skills).
• Functional domain or industry.
• Certifications explicitly requested.
• Real synonyms you also use (CRM versus HubSpot when accurate).

A 20-minute method: from posting to CV

Step A: paste the posting into a document and highlight 8–15 terms. Step B: sort them into three buckets — ones you have and already show, ones you have but do not show, and ones you do not have. Step C: add only bucket two with context. Step D: decide what to do with bucket three (learn, omit, or mention a close transferable honestly).

Example. Posting: CSM, Salesforce, onboarding, churn, NPS, QBR. Your reality: CS work in HubSpot, measured satisfaction, ran quarterly reviews, never used Salesforce. Bucket 1: onboarding, churn if you reduced it. Bucket 2: NPS or satisfaction, QBR as quarterly customer reviews. Bucket 3: Salesforce as CRM HubSpot; Salesforce in progress only if true.

Where to place them (impact order)

First, professional summary: 2–4 lines with role, context, and 2–4 anchor keywords. Second, experience bullets: each key keyword should appear at least once inside an achievement. Third, skills section: a scannable list of true tools. Fourth, education or certifications if relevant. Avoid a final block of loose keywords: it adds little and smells like stuffing.

Before (stuffing): Keywords Salesforce Salesforce Salesforce customer success churn churn NPS. After: Customer Success in B2B SaaS. Cut churn 12% by running onboarding and QBRs in Salesforce; tracked NPS and expansion. Less repetition, more credibility, same useful matches.

Synonyms and variants: when to duplicate

Sometimes the posting says MS Excel and you wrote Excel. Or JS versus JavaScript. Include the posting form at least once if it is truly your tool. You do not need to spam both ten times. For titles, use your official company title and, if the market uses another name, clarify honestly.

Watch false synonyms. Customer support is not automatically Customer Success. If your work was reactive tickets, do not sell it as strategic CS. You can show transfer (support plus product adoption) without inflating the title.

Density: how many keywords are enough

There is no magic number. A practical guide: cover the 5–8 hard requirements in the posting that are true, spread across summary, experience, and skills. If a tool already appears in an achievement, you do not need it in every bullet. Extra repetition does not replace results.

If you still get no interviews after aligning keywords, also audit ATS format (selectable text, sections, dates). Keywords will not save an image-only PDF.

Mistakes that backfire (or make you look bad)

• White text or invisible font keyword lists.
• Copying whole paragraphs from the posting.
• Inventing tools because they are trendy.
• Listing 40 skills when you use six.
• A CV keyword claim that contradicts LinkedIn.

The cost of lying is not only ethical: technical interviews and practical tests expose it. A shorter true CV beats a fragile optimized one.

Full example: before / after

Before (bullet): Responsible for customer satisfaction and the CRM. Weak keywords, no magnitude. After: Managed 80 accounts in HubSpot, launched a new onboarding, and raised CSAT from 3.6 to 4.4 in two quarters. If the posting wants HubSpot, onboarding, and CSAT, you now have match plus evidence.

Before (skills): icons and bars. After: HubSpot, Excel (Power Query), Looker, Zendesk, B2B onboarding. The ATS indexes; humans prioritize.

Keyword checklist before you apply

• Did I extract 8–15 terms from the posting?
• Did I mark which ones are true in my experience?
• Does the summary include 2–4 role anchors?
• Do critical tools appear in achievements or skills as text?
• Did I remove stuffing, white text, and absurd lists?
• Do claims match LinkedIn and what I can defend live?

How ClarCV helps without pushing stuffing

Upload your PDF or Word and check whether skills and achievements read clearly. Regenerate with an ATS template so keywords exist as selectable text, not icons. Start free; Pro is €7/month. You decide which terms are true: the product should not invent your stack.

Pair this article with how to make a CV for ATS filters and with tailoring a CV to a job ad: keywords plus structure plus evidence is the trio that moves the needle most.

Before you send, ask someone outside your field to read the CV for twenty seconds and say what you do. If they cannot answer, the summary or first role needs more clarity. That human test complements ATS checks and prevents surprises in real processes.

Keep one stable base version and create variants only when the posting changes role family. That reduces date errors and keeps consistency across applications. ClarCV helps regenerate clean variants without rebuilding layout by hand every week.

If your path has gaps, do not hide them with confusing dates: parsers and recruiters notice. Explain the period briefly in one line or prepare for the interview. Clarity usually builds more trust than cosmetic calendar cover-ups.

Also check consistency with LinkedIn: same titles, employers, and approximate dates. A large mismatch creates unnecessary doubt even when the CV parses well and keywords match the posting.

Prefer evidence over adjectives. Replace excellent communicator with an achievement that proves communication: client presentations, docs the team adopted, or fewer tickets after better internal guides.

Before you send, ask someone outside your field to read the CV for twenty seconds and say what you do. If they cannot answer, the summary or first role needs more clarity. That human test complements ATS checks and prevents surprises in real processes.

Keep one stable base version and create variants only when the posting changes role family. That reduces date errors and keeps consistency across applications. ClarCV helps regenerate clean variants without rebuilding layout by hand every week.

If your path has gaps, do not hide them with confusing dates: parsers and recruiters notice. Explain the period briefly in one line or prepare for the interview. Clarity usually builds more trust than cosmetic calendar cover-ups.

Also check consistency with LinkedIn: same titles, employers, and approximate dates. A large mismatch creates unnecessary doubt even when the CV parses well and keywords match the posting.

Prefer evidence over adjectives. Replace excellent communicator with an achievement that proves communication: client presentations, docs the team adopted, or fewer tickets after better internal guides.

Before you send, ask someone outside your field to read the CV for twenty seconds and say what you do. If they cannot answer, the summary or first role needs more clarity. That human test complements ATS checks and prevents surprises in real processes.