ATS CV Parsing Guide for Recruitment Agencies

If you've worked in recruitment for more than a week, you've seen ATS parsing failures: a well-formatted resume goes in, and a jumbled mess of data comes out. Employment dates land in the skills section. Job titles appear where company names should be. Whole sections of qualifications go missing.

This isn't random. Most applicant tracking systems were built for simple, single-column text documents, and their parsing engines haven't kept pace with modern resume design.

Parsing vs. reading

A human reader recognizes patterns instantly: the large text at the top is a name, dates on the right line up with the position on the left, a section titled "Education" contains academic history even if it's formatted differently from the last resume you read.

An ATS doesn't read — it parses. It applies fixed rules to extract data from expected locations in a document. When a resume doesn't match those expectations, which creative layouts often don't, the parsing fails.

Tables

Tables are one of the most common causes of parsing failures. They look clean to a human reader, but most parsers read text left-to-right, top-to-bottom, regardless of visual layout. A two-column table with dates on the left and job details on the right becomes a jumbled string: "2020 Senior Marketing Manager 2022 Acme Corporation Led team of 12..." Dates, titles, companies, and descriptions all merge together, and the parser either discards or misassigns the data.

PDF text layers

Modern PDFs can store text in different ways: as actual selectable text, as vector shapes that draw each letter, or as flattened images with no text layer at all. When a candidate exports a resume from Canva, InDesign, or certain Mac apps, the result often has no parseable text. It looks perfect on screen but is invisible to an ATS.

Fonts

Fonts embed into PDFs in different ways: fully, partially (only the characters used in the document), or not at all (referencing a system font instead). When an ATS encounters a font it can't render, it may substitute a font and scramble character mapping, extract text incorrectly, or fail outright. Decorative or unusual fonts are especially likely to cause problems.

How major ATS platforms differ

Not every ATS fails the same way, and knowing the specific behavior of the systems your clients use improves your success rate.

Bullhorn

Widely used in staffing, and strict about formatting. Tables are essentially unsupported, multi-column layouts fail consistently, and header/footer content is often ignored. Single-column, standard-font CVs work best.

Greenhouse

Popular with tech companies and startups. Its parser prioritizes structured data extraction and expects conventional section headers — "My Journey" instead of "Work Experience" may not parse at all.

Workday

Known among job seekers for aggressive, unforgiving parsing. Plain formatting, standard section headers, no graphics, no tables, no columns is the safest approach.

iCIMS

Reasonably robust, but still struggles with tables, columns, and non-standard fonts. Very long CVs can also truncate silently, losing experience from senior candidates.

Newer platforms: Lever, Ashby, SmartRecruiters

These have invested more in parsing creative formats, but still struggle with text-as-image PDFs and complex layouts. More forgiving of minor formatting variations, not immune to design that pushes too far.

Reformatting instead of rejecting

The answer isn't to reject every creative CV — many strong candidates invest in resume design. It's reformatting that preserves the content while fixing the structure: dates stay with their positions, skills stay categorized, contact information appears where the system expects it.

That's what Distill's reformatting engine does — it reads the document at the structural level, regardless of original formatting, and reconstructs it into a format that parses correctly across major systems. You can also target a specific platform if you're submitting high volume to one client.

Why this matters beyond parsing

When a hiring manager opens a resume from their ATS, what they see is the parsed result, not the original file. If parsing failed, they see disorganized data, even if the source resume was well designed. Clean parsing means the candidate is evaluated on the qualifications they actually have.

The guides in this section cover individual ATS platforms in detail — common failure modes and exact fixes for each.