ATS resume guide: how applicant tracking systems read your CV

An applicant tracking system (ATS) is the software companies use to collect applications, store resumes and move candidates through hiring stages. Workday, Greenhouse, Lever, SAP SuccessFactors, iCIMS and Taleo are common examples. Smaller companies often use the hiring module of their HR software, and job boards run similar parsing when you upload a resume to your profile.

You will often read that ATS software automatically rejects most resumes before a person sees them. The claim is repeated everywhere, usually without a source you can check, and it gives the wrong picture of how these systems work. The real risks are more ordinary, and you can fix them yourself.

What an ATS actually does

Most systems do four things with your application:

1. Parse your file. The software extracts the text and tries to split it into fields: name, email, phone, job titles, employers, dates, education, skills.
2. Store a candidate profile. The parsed fields go into a database record, usually with your original file attached.
3. Let recruiters search and filter. A recruiter types “product manager”, “Python” or “Berlin”, or filters by location, degree or years of experience, and gets a list of matching candidates.
4. In some setups, rank or match. The system scores candidates against the job description using keywords and parsed data. How much weight a recruiter gives that score varies a lot between companies.

Automatic rejection does happen, but mostly in the application form. Many employers add knock-out questions: “Do you have the right to work in the UK?”, “Do you hold a valid nursing license?”, “Are you willing to work night shifts?”. Pick the wrong option and the system can move you to “rejected” without anyone opening your file.

For the resume itself, the risk is quieter. If parsing goes wrong, your profile ends up with missing jobs, broken dates or skills glued into the wrong place. A recruiter searching for your skills may not find you, or opens a profile that looks careless even though your PDF looked fine. Here is what a badly parsed work history can look like inside the system:

Job title: Python SQL Tableau Senior Analyst

Employer: (empty)

Dates: 2023

The sidebar skills were read on the same line as the job title, the employer sat in a text box the parser skipped, and the dates were split across two lines. The PDF looked perfect.

Answering knock-out questions

• Read each question literally. “Do you have 3+ years of experience with Salesforce?” asks about Salesforce, not CRMs in general. If the honest answer is no, say no; a “yes” you cannot back up comes out in the first interview.
• Right-to-work and sponsorship questions are about the country of the job, not your passport in general. “Will you now or in the future require sponsorship?” means: will this employer have to sponsor a visa for you to do this job?
• Numeric fields (years of experience, salary expectation) are often filtered on. Enter a real figure. Typing 0 or 1 to dodge a salary field can drop you out of a range filter, and a recruiter who sees it will read it as a skipped question. Research the going rate for the role and city first.
• Keep a document with your answers to the questions that come up again and again (notice period, earliest start date, salary range, work authorisation) so you answer them the same way on every portal.

What parsing extracts, and where it goes wrong

Field	What the parser looks for	Typical failure
Name and contact	Text near the top: name, email, phone, city, links	Contact details placed in a Word header or inside an image
Work history	Job title, employer, location, start and end dates, description	Dates in odd formats, title and employer on lines the parser cannot pair
Education	Degree, school, graduation date	Degree written only as an abbreviation the system does not know
Skills	Words under a “Skills” heading, plus terms found in descriptions	Skills shown as icons or rating bars with no text
Languages and certifications	Named sections with plain text	Logos of certificates instead of their names

You cannot see any of this by looking at the page. The parser reads the text layer of the file, not the picture, so a resume can look perfect and still parse badly.

Myth vs fact

Myth	What actually happens
“The ATS rejects most resumes automatically.”	Most rejections are decided by people, or by knock-out questions in the application form. The resume mainly needs to parse cleanly and be findable.
“You need a special ATS template.”	Any template with real text, standard headings and a sensible reading order works. Plainness alone does not make a resume readable.
“Two columns always break the ATS.”	What matters is the order of the text inside the file. A well-built two-column PDF parses fine; a badly built one does not.
“PDFs cannot be read by ATS.”	Text-based PDFs are read well by current systems. Scanned PDFs and images are the problem.
“Hide keywords in white text to beat the filter.”	The hidden text shows up in the parsed profile, where recruiters can see it. It looks dishonest.
“More keywords means a higher score.”	A list of keywords with no context is easy to spot and convinces nobody. Keywords work best inside real achievements.

File formats

• Text-based PDF: the best default. It keeps your design, and current systems extract the text well. Most resume builders and word processors export this.
• Word (.docx): fine when the job ad asks for it, and some older systems handle it slightly better. Layouts built with text boxes or Word headers can still go wrong.
• Scanned or image PDF: avoid it. There is no text to extract, so the parser gets nothing or relies on character recognition that mangles names and dates.
• Apple Pages (.pages), .odt, images (.jpg, .png): many systems do not accept them, and those that do may not parse them. Export to PDF.
• Google Docs: download it with File → Download → PDF or Word. Do not send a share link unless the employer asks for one.

If the portal lists accepted formats, use one of those. When it accepts both PDF and Word and says nothing else, send the PDF.

Layout: what helps and what hurts

Columns and reading order

A parser reads the text in the order the file stores it. In a good file, that order is name, contact details, profile, experience, then the sidebar. In a bad one, the text jumps between columns line by line, and your skills get merged into a job description.

Two-column resumes are therefore fine when the file is built properly. AlderCV renders every template from the same HTML structure in a logical order (header, main content, sidebar) and automatically checks the extracted text of generated PDFs, so the order you see is the order a parser gets. If you are applying to a very conservative employer or through an old portal, a single-column template like Linden or Meridian settles the question.

Tables and text boxes

Tables used for layout are read cell by cell, and the order depends on how the table was built. Text boxes in Word are sometimes skipped. Keep your contact details and your work history out of both.

Headers and footers

Some parsers ignore the header and footer areas of a Word document. Put your name and contact details in the body of the page, at the top. A page number in the footer is harmless.

Icons, graphics and skill bars

A phone icon is not a phone number. Icons are fine as decoration next to the text they stand for. Skill bars, stars and pie charts carry no text, so the parser sees nothing. Write the level in words if it matters: “German (B2)”, “SQL, advanced”.

Fonts

Common fonts extract cleanly. Some decorative fonts and ligatures turn into odd characters when the text is copied out (“finance” instead of “finance”, missing letters in names). If your name or employers contain letters such as ü, ş or é, check that they come through intact.

Links

Many parsers keep only the visible text, not the link behind it. Write the address itself, “linkedin.com/in/anna-meyer”, instead of a word like “LinkedIn” that hides the URL.

Section headings parsers recognise

Parsers find sections by their titles. Creative headings make them guess.

Use	Instead of
Profile or Summary	About me, Who I am
Experience or Work Experience	Where I’ve been, My journey
Education	Academic background story
Skills	My toolbox, Superpowers
Languages	Words I speak
Certifications	Badges
Projects	Things I built

In AlderCV the default section titles follow your resume’s language, so an English resume gets “Experience” and “Education”, and a German or Turkish one gets the standard local headings.

Dates

Use one format throughout, with month and year for recent roles: “Mar 2022 – Present” or “03/2022 – Present”. Avoid seasons (“Summer 2021”), and avoid putting the year alone on one line and the role on another. Parsers calculate years of experience from these dates, and a recruiter filtering for “5+ years” depends on that calculation.

If you were promoted or changed roles inside one company, give each title its own dates under the employer:

Northwind Logistics, Leeds

Senior Analyst, Jan 2023 – Present

Analyst, Sep 2020 – Dec 2022

If the autofill test (below) then loses the older title or merges the two, repeat the employer name on each role instead. Our guide to showing a promotion on a resume has both layouts with bullets and the edge cases.

Keywords: matching the job ad without stuffing

When recruiters search, they type words from the job ad. Some systems also compare your resume with the ad automatically. Either way, the words you use decide whether you turn up.

A worked example

Say the ad for a Marketing Operations Specialist reads, in part:

You will own our HubSpot setup, build lead scoring and nurture workflows, and report on funnel performance in Looker Studio. Experience with Salesforce integration and A/B testing is a plus.

The keywords worth noting are HubSpot, lead scoring, nurture workflows, funnel reporting, Looker Studio, Salesforce, A/B testing, and the job title itself.

Now compare a bullet from a resume that has done this work but describes it vaguely:

Before: Managed the marketing tools and helped the sales team get better leads.

After: Rebuilt lead scoring and three nurture workflows in HubSpot, synced with Salesforce, which raised the share of leads sales accepted from 31% to 44%. Reported funnel performance weekly in Looker Studio.

The second version names the tools the ad names, describes the same work in the ad’s language and adds a result. A recruiter searching “HubSpot lead scoring” now finds it, and the person reading it learns something.

Some rules for keywords:

• Use the exact job title where it is true (“Customer Success Manager”, not “client person”).
• If your official title was unusual, write the standard equivalent and keep the real one in brackets: “Customer Support Specialist (internal title: Customer Happiness Hero)”. Recruiters search for the first, and a reference check will confirm the second.
• Spell out abbreviations once: “search engine optimisation (SEO)”. Some recruiters search the long form, some the short one.
• Put keywords inside bullets and your profile. A skills section helps, but a skill that appears nowhere in your experience looks unproven.
• Only use terms that describe your real work. If you have never touched Salesforce, leave it out and be ready to say you are keen to learn it.

In the AlderCV editor, Match a job does this comparison for you: paste the ad and it shows which of its key terms your resume already uses and which are missing, and the count updates as you edit. It is free on every plan. With Pro or the 30-day pass, the AI match adds a score and suggestions tied to your entries; nothing changes until you accept one. The full process, from reading the ad to adjusting bullets, is in our guide on tailoring a resume to a job description.

How to test your own resume

1. The copy-paste test. Open your PDF, select everything (Ctrl/Cmd + A), copy it and paste into a plain text editor. If the text comes out complete, in a sensible order, with your name at the top and dates next to the right jobs, a parser will read it too.
2. The autofill test. Many application portals offer “upload your resume to fill in the form”. Upload yours on a real application and look at what lands in each field before you correct anything. Missing jobs or swapped dates show you exactly where the file confuses parsers.
3. The search test. Pick the five most important terms from the ad. Search your resume (Ctrl/Cmd + F) for each one. If one is missing and you genuinely have that skill, add it where you used it.
4. The ATS checker. Our free ATS resume checker runs the copy-paste test for you and goes further: it flags columns that get mixed line by line, spaced-out headings, missing contact details and inconsistent dates, and it can compare your resume with a job ad. The file is read in your browser and never uploaded.
5. The resume check. If you write in AlderCV, run the resume check in the editor before you download; it points out problems that are easy to miss on your own.

FAQ

Do I need a special “ATS template”?

No. Any template with real text, standard headings and a logical text order works. Many templates sold as “ATS-friendly” are simply plain. A plain resume with a scrambled text layer is worse than a designed one with a clean one.

Can an ATS read a two-column resume?

Yes, if the text inside the PDF is stored in a sensible order. Run the copy-paste test to see what a parser sees. If the result jumps between columns, switch templates.

Should I add a separate keyword section?

A normal skills section already does this job. A block of keywords with no context, or keywords hidden in white text, adds nothing a recruiter trusts. Read our skills section guide for how to build one that works.

Does a photo break the ATS?

A photo is an image, so the parser ignores it and reads the rest of the file as usual. Whether to include one is a question of local norms, covered in our guide to resume photos by country.

PDF or Word?

A text-based PDF, unless the ad or the portal asks for Word. Name it clearly: First-Last-CV.pdf. When Word is required, “Download as” in the AlderCV editor’s “⋯” menu gives you a plain, single-column .docx of the same resume; the design stays in the PDF. Our guide to Word or PDF covers when to send which, and how to fill in application forms.

Does the ATS read my cover letter?

Usually it stores it as a file attached to your profile. People read it, so the advice in our cover letter guide matters more than any parsing rule.

ATS checklist

• Standard section titles: Profile, Experience, Education, Skills, Languages
• Name and contact details as real text at the top of the page, not in a Word header
• Dates in one consistent format (e.g. "Mar 2022 – Present"), each title with its own dates
• No important information in images, tables, text boxes or rating bars
• Icons always next to the text they stand for; links written out
• Text-based PDF; the copy-paste test produces a clean, ordered result
• Job title and key skills match the ad where it is honest, inside real bullets
• Knock-out questions on the application form answered literally and truthfully, with real figures in numeric fields
• A sensible file name: First-Last-CV.pdf

How AlderCV helps

• Every template uses real text, standard default section titles and clickable links that show the address.
• Icons are decorative and always sit next to the text.
• PDFs are generated with a logical reading order and document structure, include title and author metadata, and the extracted text order is tested automatically.
• Section titles follow your resume’s language.

For the writing itself, start with how to write a resume that gets read and the list of resume action verbs.