How do I make my resume ATS-friendly so it gets past the applicant tracking system?

An ATS-friendly resume is one that an applicant tracking system can read and file correctly: the software extracts your text, sorts it into fields like name, work history, education, and skills, and stores it so a recruiter can search and rank it [1][6]. Making a resume ATS-friendly is mostly a matter of clean structure and relevant wording, not of defeating a gatekeeper. Almost every large employer runs this software. In a survey of 384 recruiters, the resume-tools company Jobscan found that 99.7 percent of them use an ATS, and the Harvard Business School and Accenture report Hidden Workers: Untapped Talent noted that 99 percent of Fortune 500 firms used applicant tracking software as of 2019 [2][6].

What the system is actually doing

When you upload a file, the parser converts it to plain text, strips the visual layout, and tries to map each piece to a field in a candidate profile [1]. Named entity recognition tags parts of the text as a person, an organization, a date, a job title, or a skill, then writes them into the record a recruiter reads [1]. On many systems that parsed profile, not your original design, is the first thing a person sees, so a clean parse decides whether your experience lands in the right place [1].

What genuinely helps

Most of the payoff comes from getting a few ordinary things right. Structure the document so the parser can follow it, which means a single reading column, plain section names, and contact details kept in the body of the page rather than a header [2][3][5]. Use the real language of the posting, because the system indexes your words and lets recruiters search for the skills and titles a role calls for [2][3]. And submit a file the parser can open as text, a .docx or an exported PDF rather than a scan or an image [4][5]. The sections below take each of those in turn, then separate the genuine risks from the folklore.

Use a single-column layout with plain, conventional section headings, a common font, and your contact details in the body of the page [2][3][5].

Keep the resume to a single reading column

A parser reads a page as one stream of text from top to bottom, so a two-column design forces it either to read straight across both columns or to take one column entirely before the other, which jumbles your dates and job titles [2][3]. Indeed notes that tables, columns, headers, and footers can scatter or lose information when the system translates the page [3]. A single column, with each job in date order, keeps the reading order intact. Save graphical touches for a version you send straight to a person or share as a portfolio.

Use section names the parser expects

Name your sections with terms a parser is trained to recognize, such as Work Experience, Education, and Skills, because the system uses those headings to tell where one section ends and the next begins [1][3]. A creative label like My Journey or What I Bring can leave the parser unsure where your jobs are, so it may file the content under the wrong field or miss it [2][3]. Indeed points out that a system might be set to look for the word Skills but not Abilities, which is why the ordinary word is the safer choice [3]. The University of Pittsburgh career center gives the same advice, suggesting you keep the word Experience in your experience headings [5].

Pick a common font and keep contact details in the body

Choose a widely installed font such as Arial, Calibri, or Times New Roman at about 10 to 12 points, and type your name, phone, and email into the main body rather than the header or footer [2][5]. A parser that does not have your font falls back to a default and can garble spacing, and many systems skip header and footer layers, so details placed there can vanish [2][5]. Jobscan notes that custom fonts and clever icons often parse as gibberish or a [NULL] tag, and it recommends plain round or square bullets over arrows and checkmarks [2].

Take the specific skills, tools, and job titles named in the posting and work them into your resume in plain, truthful language, because the system indexes your text and lets recruiters search and rank candidates by those terms [2][3].

Read the posting closely and note the words it repeats, since a term that appears several times is likely one the employer entered into the system as a search or ranking criterion [3][5]. Put those terms where they are true: in your job bullets, a short skills line, and your titles, rather than piled into a separate keyword block detached from your experience. Include both the spelled-out phrase and its acronym at least once, for example Search Engine Optimization (SEO), because a given system may search for one form and not the other [3]. Match the employer's phrasing where it fits your real experience; if the posting says customer success and your resume says client relations, adding the posting's exact term helps the search find you [2].

The reason to use the job's own language is not that a machine grades your vocabulary. It is that human reviewers configure filters and searches around the criteria in the description. The Harvard Business School and Accenture report Hidden Workers found that more than 90 percent of employers who use this software rely on it to make a first cut, and it estimated that 27 million qualified people in the United States are screened out of roles they could do, largely because employers configure filters around exact-match criteria [6]. Mirroring the posting's real requirements is how you stay on the right side of that filter without misrepresenting yourself.

Either a text-based PDF or a Word .docx parses well on current systems, so the safest rule is to send whatever the job posting asks for and to make sure the file holds real selectable text rather than a scanned image [4][5].

Modern parsers read both formats, though there are small differences. Jobscan generally recommends a PDF because it holds your layout together, and switches to .docx only when a posting specifically requests Word; it also notes that some systems have trouble with special characters in .docx files [4]. Whichever you pick, keep both versions saved so you can match a posting's requirement quickly [4]. The format that reliably fails is a scanned or photographed resume saved as an image inside a PDF, because it carries no text for the parser to extract [1][4]. When you make the PDF, export or save it directly from your writing app; the University of Pittsburgh career center specifically warns against a print-to-PDF path, which can produce a file the parser reads poorly [5]. If an online application gives you a plain-text box to paste into, paste a clean text version and fix any spacing the paste introduces [2].

The formatting that does the most damage is the kind that fails without any visible error, mostly text hidden inside images and objects the parser cannot open [1][2].

Anything stored as a picture is invisible to a text parser, so a name inside a designed banner, a logo, or a skills graphic simply never reaches your candidate profile [1][2]. Greenhouse lists resumes with graphics, photos, or word art, and files uploaded as an image rather than a .docx or PDF document, among the common reasons a resume fails to parse; it also cannot read a file larger than 2.5 MB [1]. Skill-rating bars are a frequent trap: a graphic showing a language at four of five dots carries no readable text, so replace it with something like Java (advanced) [2]. Tables and text boxes cause similar trouble, because some systems read a table out of order or treat a text box as a separate object and skip it [1][3]. A quick self-check is to open your saved file, select all, copy, and paste it into a plain text editor; whatever comes through in a sensible order is roughly what the parser sees, and anything missing or out of place is a trap to fix [2].

The biggest myth is that a bot automatically rejects most resumes before any person sees them, and the widely repeated figure behind it does not hold up [7][8].

The 75 percent auto-rejection claim

The claim that roughly 75 percent of resumes are auto-rejected by software traces back to a 2012 sales pitch from a company called Preptel, which sold resume-optimization services and went out of business the next year without publishing any methodology [7]. Career analysts who went looking for the underlying study found none; the number was repeated so often that it began to sound like research [7]. Applicant tracking systems weigh, sort, and filter, but people do the reviewing [8], and recruiters interviewed for one analysis estimated that a person still looks at the large majority of applications [7].

Hidden white-text keywords

Typing extra keywords in white or tiny font to hide them from a human reader does not reliably beat the system, and it tends to backfire. The National Resume Writers' Association notes that the applicant tracking system reads the white text and makes it visible on the parsed resume, so a recruiter sees the hidden block and reads it as an attempt to deceive [9]. Its writers compare the tactic to cheating on a test, on the grounds that you will be found out [9]. Keywords woven honestly into your accomplishments do the same indexing job without the risk [9].

PDFs cannot be read

The idea that any PDF is unreadable is out of date. Current systems parse text-based PDFs well, and the real failure case is a scanned or image-only PDF with no selectable text [4]. If you can highlight and copy the words in your PDF, a parser generally can read them too [4].

An applicant tracking system does not, on its own, decide that you are unqualified and delete you. It stores, parses, ranks, and searches, while people set the rules and make the calls [1][8].

It does not judge visual design. The parser reduces your resume to text and fields, so a polished template earns you nothing with the software and can cost you if the design hides text in images or columns [1][2]. It does not apply one universal standard either, because vendors such as Greenhouse, Workday, iCIMS, and Taleo each parse and rank differently, which is why a plain, conventional resume that reads cleanly everywhere beats one tuned to a single system [1][2]. It does not usually reject candidates automatically; when an application is knocked out without review, that is normally a screening question an employer configured, such as work authorization or a required credential, rather than the parser exercising judgment [7][8]. HiringThing puts it plainly: the filters are set by people using the software, not by the software itself, so the outcomes reflect an employer's hiring choices [8]. And it does not replace the recruiter. The system's job is to make your experience findable and readable, but a person still decides whether to move you forward [7][8].