Resume guides

ATS-friendly résumés: what actually matters

Applicant tracking systems are database tools, not judges. They rarely reject a candidate on their own — but a document they cannot parse arrives at the recruiter's screen as garbled fields and missing dates.

What an ATS actually does

• Extracts text from your file and maps it to fields: name, employer, title, dates, education, skills.
• Stores the result so recruiters can search the candidate pool by keyword.
• Ranks or filters candidates only when a recruiter configures it to.

So the goal is not to trick a machine. The goal is a document that parses cleanly and contains the words a recruiter would search for.

Formatting rules that keep parsing clean

• Use a single-column layout. Two-column designs frequently interleave text during extraction.
• Put contact details in the body, never in a header or footer element.
• Avoid text inside images, icons, text boxes and charts — none of it is extracted.
• Use standard section headings: Experience, Education, Skills. Creative labels like 'My journey' are not recognised.
• Use a common font and plain bullet characters.
• Write dates consistently, e.g. 03/2021 – 08/2024.
• Submit PDF unless the posting asks for .docx.

Keywords without keyword stuffing

Take the job posting and list the nouns: tools, certifications, methodologies, domains. Those are the searchable terms. Include each one where it is true, inside a real sentence about real work.

Hidden text and other tricks

White-on-white keywords, invisible fonts and prompt-injection lines aimed at AI screeners are detected and treated as fraud by most employers. The downside is disqualification and a blacklist entry; there is no upside.

How parsing actually fails: a worked example

Consider a résumé built in a two-column template: a narrow left column with contact details and skills, a wide right column with experience. Visually it looks organised. When an ATS extracts text line by line, it often reads left-to-right across the page rather than down each column, so a line from the skills column merges with a line from the experience column mid-sentence. The employer name and the job title can end up on separate, disconnected fields, and the parser may store your phone number as part of your most recent job title.

The fix is not a workaround inside the two-column design; it is a single column. If you want a compact, designed look for a human reader, keep a plain single-column version as the file you actually submit through online application forms, and reserve any more visual layout for a printed copy you hand over in person.

Skills sections: matching without padding

1. Copy every tool, certification and methodology named in the posting into a scratch list.
2. Cross out anything you cannot honestly claim to have used in real work.
3. Group what remains under two or three headings, such as Tools, Methods, Languages.
4. Fold the most important two or three into a bullet in your experience section as well, so the term appears in context, not only in a list.

Common ATS mistakes beyond formatting

• Uploading a scanned image of a printed Résumé instead of a text-based file — there is no text for the system to extract at all.
• Naming the file something generic like 'resume_final_v3.pdf', which is harmless for parsing but looks careless once it reaches a recruiter's downloads folder.
• Using a table to lay out the Skills section — tables are parsed inconsistently across systems, so a plain comma-separated list is safer.
• Listing job titles only in a separate 'Skills matrix' at the top without also stating them plainly next to each employer in the experience section.
• Submitting the same file for every application when the posting uses noticeably different terminology than your current draft.

A two-minute self-test

1. Open your PDF, select all, copy, paste into a plain text editor.
2. Read the result. If the order is scrambled or fields are missing, the ATS sees the same thing.
3. Check that employer names, titles and dates each survived on their own line.

Different systems, different quirks

Not every applicant tracking system behaves the same way. Some rank candidates automatically against the posting; others simply store the parsed fields for a recruiter to search manually with no ranking at all. A few older systems still struggle with anything beyond a single font and basic bullet points, while newer ones tolerate light formatting fairly well. Since you rarely know which system a given employer runs, the safest approach is to design for the strictest case every time rather than gambling on the more forgiving one.

Building a keyword list without guessing

1. Paste three to five postings for the same role title into one document.
2. Highlight every noun that repeats across at least two of them — that is your core list.
3. Separate the list into 'have and can prove' and 'have some exposure to'; only the first group belongs in bullet points, the second can sit in a skills line.
4. Check your current draft against the list and note anything missing that is actually true of your experience.
5. Leave out anything you cannot speak to in detail if asked about it directly in an interview.

A note on PDF versus Word

Most modern systems parse PDF correctly, but a PDF built by exporting a heavily designed document — with layered text boxes or an embedded image of a logo behind the header — can still confuse a parser even though it looks fine to a human. If a posting explicitly requests a .docx file, follow that instruction rather than defaulting to PDF; it usually means their specific system has had documented trouble with PDFs before.

When you suspect the system rejected you unfairly

If you are confident you match a posting well and hear nothing back within the usual window, it is reasonable to look for a referral or a direct message to someone on the team rather than assuming the formatting is the sole cause — most silent rejections are about fit or volume of applicants, not parsing. Fixing formatting is worth doing once, well, and then moving your energy to targeting and outreach.