How ATS reads resumes (and what to fix first)

What an ATS actually does

Think of an ATS as a pipeline, not a judge of your career. In a typical flow it stores your file, extracts text, normalizes that text, and then searches or scores it against the role. Recruiters may later filter by keywords, years of experience, location, or education — but none of that works well if the raw text extraction is messy.

Parsing is the fragile step. The system has to turn a visual document into a sequence of characters and, ideally, recognize sections such as Experience, Education, and Skills. When the layout is decorative rather than structural, the extracted text can scramble order, drop content, or glue unrelated words together.

Keyword matching usually happens on that extracted text. If a skill appears only inside an image, a chart, or a text box the parser skips, it may never enter the searchable record — even though a human looking at the designed PDF can see it clearly.

How resume files get turned into text

PDF and DOCX are the common upload formats. DOCX is often more predictable for text extraction because it is a structured document format. PDFs vary widely: some are clean text PDFs; others are visual layouts, exports from design tools, or scans.

A clean text PDF still has a reading order. Multi-column layouts, sidebars, and tables can cause the extractor to read left column then right column — or worse, line-by-line across columns — so a job title lands next to dates from another role.

Scanned pages and image-only PDFs are the hardest case. If there is no real text layer, an ATS may store almost nothing useful unless optical character recognition (OCR) is applied, and OCR quality varies. If you can select and copy text in your PDF viewer, you are in better shape than if the page behaves like a picture.

• Prefer a selectable-text PDF or a DOCX over a scanned image.
• Avoid locking critical skills inside logos, icons, or skill “bars.”
• Keep a simple top-to-bottom reading order whenever you can.

Formatting mistakes that break parsing

These patterns look polished in a design portfolio and often fail in automated intake. You do not need a boring resume — you need a readable one.

• Tables used as the whole page layout: many parsers read cell-by-cell and destroy chronological order.
• Multi-column designs and sidebars: contact details or skills in a narrow column may be read at the wrong time or missed.
• Text inside graphics, charts, or icons: if it is not real text, it may never be indexed.
• Headers and footers for email/phone: some systems ignore repeated header/footer regions.
• Text boxes and floating frames from design tools: extraction order becomes unpredictable.
• Unusual section titles only (“My journey”, “Wins”): humans get it; keyword filters may look for Experience, Education, Skills.

What helps ATS — and humans — read you clearly

Clear section headings, consistent date formats, and role blocks that follow a predictable pattern (title, company, dates, bullets) help both software and recruiters. Use the language of the job posting where it is truthful: if the posting says “project management” and your CV only says “ran initiatives,” you may be describing the same work in words the filter never searches for.

Bullet points that start with strong verbs and include concrete outcomes are easier to scan than dense paragraphs. Metrics help when they are real; invented numbers hurt trust later in the process.

Length is secondary to clarity. A two-page specialist CV can parse fine. A one-page CV full of columns and icons can parse poorly. Optimize for extraction first, then tighten wording.

Cover letters in an ATS world

Many employers still collect a cover letter in the same portal. Treat it as searchable text too: plain paragraphs, no text locked in a letterhead image, and a direct match to the role’s requirements in the opening lines.

A short letter that names the role, mirrors a few must-have skills from the posting, and points to one relevant achievement usually travels better than a generic essay. If the portal has separate fields, paste plain text rather than uploading a heavily designed letter PDF when you have the choice.

A practical checklist before you apply

Use this as a pre-flight list, not as superstition. Fixing these issues will not guarantee an interview — it improves the odds that your content is present for whatever matching step comes next.

• Open your file and confirm you can select/copy the body text.
• Export a simple single-column layout with standard headings.
• Move skills out of graphics and into a real Skills section.
• Align wording with the job posting where it is accurate.
• Remove tables used only for visual layout.
• Proof the plain-text paste of your CV into a blank document — if that paste looks scrambled, an ATS may see the same mess.

Where ResumeTune fits

ResumeTune’s free check is a deterministic technical pass: it looks at readability and structure signals for English and Spanish CVs, and can compare keywords when you paste a job description. It does not call an AI model on the free path, and it does not claim to reproduce every vendor’s ATS ranking formula.

Use it after you simplify formatting — or use it to spot issues before you spend time rewriting. Paid options on the site are separate and optional. The point of this guide is education; the homepage upload is the tool.