ATS optimisation in 2026: what actually matters (and what doesn't)
Most ATS advice online is outdated or wrong. Here's what modern applicant tracking systems actually parse, and the three things that will get you past the filter every time.
Applicant Tracking Systems (ATS) have a reputation as black boxes that eat resumes and spit out rejections. Most of what you read about them is either a decade out of date or based on speculation. This guide separates the myths from what modern ATS platforms actually do in 2026.
What an ATS actually does
An ATS is fundamentally a database with a search interface. Recruiters post a job, candidates apply, and the ATS stores every application. When a recruiter wants to review candidates, they either scroll through the list or use a search bar to filter by keyword, location, or status.
Parsing is the first step: the ATS reads your PDF or DOCX, extracts text, and stores it in structured fields — name, email, work history, education, skills. This is where most resume "fails" happen — not because the system rejected you, but because it mangled your text and the recruiter's keyword search missed you.
Ranking is the second step, and only some ATS platforms do it. Systems like Workday, Greenhouse, and Lever focus on storage and workflow. Others, like iCIMS and Eightfold, apply ML-based matching to score candidates automatically. The advice differs slightly depending on which type you're up against — but the foundations are the same.
Note
You rarely know which ATS a company uses. Design for all of them: clean parsing first, keyword density second.
The myths
Myth: You need to stuff keywords to beat the ATS. Keyword stuffing was a real hack circa 2012. Modern ATS platforms penalise it, and even if they don't, the recruiter who reads your resume will. Focus on natural, relevant language.
Myth: Graphics and tables break all ATS systems. Some older systems can't parse multi-column layouts. Most modern systems (Greenhouse, Lever, Workday) handle them fine. The safer rule: if you're applying via a small company with an unknown stack, use single-column. If you're applying via a major platform with known ATS, you have more flexibility.
Myth: The one-page rule. Length matters much less to an ATS than it does to a human. The ATS doesn't care if your resume is 1 page or 3. Recruiters at large companies who process hundreds of applications per week do care — brevity signals self-awareness. For roles with under 10 years of experience, one page. For senior and executive roles, two pages is fine.
Myth: PDF = bad for ATS. PDFs generated from Word or Google Docs are machine-readable and parsed reliably by all major ATS platforms. The exception: scanned PDFs (images of paper resumes) — these are not searchable and will fail parsing. Always export from a proper document editor, never scan.
What actually matters in 2026
1. Exact keyword matching
Job descriptions contain specific terms. If the job says "Python" and your resume says "scripting languages", a keyword search for Python will miss you. Use the exact terms from the job description.
That doesn't mean copying the JD verbatim — it means using the terminology the employer uses:
- "Account executive" vs "sales representative"
- "Product roadmap" vs "feature planning"
- "CI/CD pipeline" vs "continuous integration"
Tip
Paste the job description and your resume into a word-frequency tool. Any high-frequency term from the JD that's missing from your resume is a gap worth filling — if it's genuinely true of your experience.
2. Clean, parseable structure
The ATS reads your resume top to bottom, left to right. Help it by using standard section headers:
- Work Experience (not "Career Journey" or "My Story")
- Education (not "Academic Background")
- Skills (not "Superpowers" or "Toolkit")
Avoid headers in text boxes, tables, or sidebars — they often parse as body text or get dropped entirely.
Use consistent date formatting: Jan 2023 – Mar 2025 or 01/2023 – 03/2025. Don't mix styles.
3. File quality
Export as a text-layer PDF from your editor. Check it by opening the PDF and trying to copy-paste a sentence — if the text is selectable and readable, you're in good shape. If it isn't, regenerate.
If the employer's portal accepts DOCX, consider submitting both: DOCX for maximum parsability, PDF as a backup if you can attach an additional file.
Section-by-section checklist
Contact information
- Full name at the top, not in the header area of a Word document template (headers often get dropped)
- Professional email — no
cooldude1987@addresses - LinkedIn URL (shortened:
linkedin.com/in/yourname) - Location: City, Country is enough. Full street address is outdated and a privacy risk
Work experience
- Reverse chronological order (most recent first)
- Company name, your title, dates, and location for each role
- 3–6 bullet points per role, starting with action verbs
- Metrics in at least half your bullets (see our guide on quantifying bullet points)
Skills section
- List technologies, tools, and methodologies explicitly — don't embed them only in bullet points
- Group by type: Languages: Python, SQL, JavaScript | Tools: dbt, Airflow, Tableau
- Avoid proficiency bars — they're subjective, not parseable, and trained recruiters ignore them
Education
- Degree, institution, graduation year
- GPA only if above 3.5 and you graduated in the last 3 years
- Relevant coursework or thesis only if directly relevant to the role
How modern ML ranking works
ATS platforms with ML ranking (Eightfold, Beamery, some Workday/SAP configurations) go beyond keyword matching. They:
- Parse semantic meaning, so "reduced churn" and "improved retention" may score similarly
- Weight recency (your last 2 roles matter more than roles from 10 years ago)
- Score title matching separately from skills matching
- May factor in trajectory — a series of promotions is a positive signal
The practical implication: tailor your resume to every job, but don't obsess over synonyms at the expense of clarity.
The 30-second test
Before submitting any application, do this:
- Copy your entire resume text and paste it into a plain text editor (Notepad, TextEdit in plain text mode)
- Read through it. Does the structure hold up? Can you parse your own work history quickly?
- Ctrl+F your top 5 target keywords from the job description. Are they all present?
If the answer to all three is yes, your resume is ATS-ready.
ResumeCommand runs this analysis automatically. When you paste a job URL, it extracts the key signals and highlights gaps between the JD and your current resume before you even start editing.
Try it free → ResumeCommand