TACTICAL INSIGHT
Why Your Résumé Can Look Perfect — and Still Be Misread by an ATS
A résumé can look correct while software extracts altered or missing characters. An observed candidate case explains how PDF text extraction can go wrong, what that may mean for ATS processing, and practical checks to make before applying.
1. The résumé problem you may never see
You open your résumé, review the formatting, check the spelling, and save it as a PDF. Everything looks right. But when software reads that same file, it may not recover the words you see on the screen.
Most of us are familiar with WYSIWYG—what you see is what you get. A PDF may look exactly as intended when you open or print it. But what you see is not necessarily what software extracts. Recovering the letters and words requires software to interpret information inside the file, and that process does not always reproduce the text displayed on the page.
A job title, skill, or qualification might appear correctly to you while a character in the extracted text is missing or replaced. The difference can be small enough to escape notice, even if the extracted text is shown to you.
That does not mean every applicant tracking system will misread the file, or that a parsing error caused a rejected application. It means your résumé could contain an issue you cannot see—one worth checking before you submit your next application.
2. What you see isn't necessarily what software reads
A PDF viewer and a résumé parser perform different tasks. The viewer uses information in the file to display letters, spacing, and formatting as intended. A parser must recover the underlying characters and words so software can process them.
Those two operations do not always produce equivalent results. The information that allows a PDF to display a letter correctly may not provide a reliable way to recover that same letter as text.
Think of the difference between recognizing a word on a printed page and recovering the exact characters stored behind it. You might see Education in the résumé. An extraction tool might recover Educa?on instead. To a person viewing the original PDF, nothing appears wrong. To software working from the extracted text, those are different strings.
The distinction matters because later processing may depend on the text that was recovered—not the page as you saw it. A résumé can therefore look finished and professional while its extracted contents deserve a closer look.
3. A real résumé that revealed the problem
We encountered this issue while examining an actual candidate's résumé. The PDF looked normal when opened, and a general-purpose AI tool could interpret its contents. Yet when we examined the document using Docent's ATS Diagnostic, the extracted text contained unexpected characters. Further investigation identified a problem with how some characters were recovered from the PDF.
For example, expertise appeared in the extracted text as experIse, and medication as medicaIon. Other affected words showed punctuation where letters belonged. In the first two examples, a lowercase letter was replaced by an uppercase character that looks similar. Without highlighting the difference, the error is easy to overlook. These were not ordinary spelling mistakes in the candidate's visible résumé. They appeared during the process of recovering text from the PDF.
The candidate had also tried a separate résumé optimization tool. It displayed the altered words as apparent misspellings, without explaining that they could be the result of a PDF text-extraction problem. That distinction is important: seeing a strange word is not the same as understanding why it appeared or knowing whether the original résumé needs to be re-exported.
Our investigation found incomplete character-mapping information associated with embedded PDF fonts. One extraction approach recovered the affected words differently from another. The case does not establish that every employer's applicant tracking system would read the résumé incorrectly, or that it caused a missed opportunity. It does establish something job seekers should know: a document can look correct while a software reading of that document contains errors the candidate never suspected.
4. Why PDF text extraction goes wrong
A PDF is designed to preserve how a document looks. It can use embedded fonts and other information to display letters, spacing, and formatting correctly. But displaying a letter and recovering that letter as text are two different operations.
Behind the visible page, a PDF contains character information that software must interpret. When that information is incomplete or interpreted incorrectly, software may extract the wrong letters even though the document looks normal. This can happen because of problems with embedded fonts, character mappings, or the way a PDF was created. The PDF standard recognizes this distinction between displaying text and extracting it. [1]
Fonts and ligatures can complicate matters further. A ligature is a typographic feature in which two or more letters, such as ti, are combined into a single character shape. In some PDFs, that combination is represented by one encoded character rather than separate letters. Extraction software must translate that character back into the correct sequence of letters. When the necessary character mapping is incomplete, the result can be missing or substituted text.
In our candidate's résumé, the combination ti repeatedly produced incorrect extracted text, appearing as ? in some words and I in others. Documented software issues involving PDF files made with Calibri and other fonts show that character-extraction problems are not unique to this case. [2][3] This does not mean those fonts are inherently unsuitable for résumés; the outcome depends on the PDF and the software interpreting it.
Our candidate's PDF illustrated the character-mapping problem directly. Its embedded font information was incomplete for certain characters, and one extraction method returned altered words while another recovered them correctly. The problem was not simply a misspelled résumé or an image-only scan. It was a mismatch between what the document displayed and what a particular reading process recovered.
5. How these problems can affect ATS processing
When an employer's applicant tracking system receives a résumé, it may use parsing technology to identify information such as contact details, job titles, work history, education, and skills. Before software can interpret those qualifications, it needs usable text—or another reliable method of recovering the information—from the submitted file.
If extraction changes medication to medicaIon, for example, a later step working with that altered text may fail to recognize the intended word. The same concern applies to a job title, certification, equipment name, or other relevant qualification. Some systems may recognize the intended word despite the error; others may not. The result depends on how each system extracts and interprets the résumé. Our case does not tell us which outcome would occur in any employer's system.
This is not just a theoretical concern about PDFs. Greenhouse, an applicant tracking system provider, documents instances where résumé parsing is unsuccessful or incomplete. Its support guidance identifies formatting problems and characters that its parsing technology does not recognize as possible causes. [4][5] That documentation does not prove Greenhouse—or any other ATS—would misread our candidate's specific file. It does confirm that parsing problems can occur in real recruiting software.
There is an important distinction between a résumé being parsed incorrectly and an application being rejected. A parsing error does not automatically disqualify a candidate, and some systems may recover information that others miss. But candidates generally have no way of knowing how each employer's software will interpret their file. If a known extraction problem changes important qualifications or experience, leaving it unresolved creates an unnecessary risk. Correcting the problem before submitting the résumé is the more reliable approach.
6. Why these errors are so easy to miss
The first obstacle is that the résumé itself may give no warning. The page opens normally. Headings line up. Words appear correctly spelled. Nothing about the visual presentation tells the candidate that a text extractor might recover something different.
The second obstacle is that even extracted text can appear mostly right. A substitution like experIse instead of expertise is easy to skim past; another error might replace letters with punctuation. In a long résumé, a few altered words can disappear into otherwise readable paragraphs. A spelling check may flag an odd word, but it cannot by itself establish whether the original file was misspelled or the extraction process changed it.
This is why a parsing problem can be mistaken for a spelling problem. A candidate may see an unusual word in an optimization report without realizing that the original résumé looks correct and the error appeared during extraction.
A more useful check compares what the PDF visibly says with the text actually extracted from it, drawing attention to suspicious differences. No single check can guarantee how every employer's system will process a résumé. But checking the extracted text can reveal problems that a visual review misses—and give candidates an opportunity to correct them before applying.
7. How to check and protect your résumé
You do not need to understand PDF technology to look for text-extraction problems. A few checks can help you catch errors before you submit an application.
-
Start with the finished file. Check the actual PDF you plan to upload, not just the Word, Pages, or other document you created it from. A problem can arise when the file is exported, even if the original looks fine.
-
Copy the text out of the PDF. Open the PDF, select its text, and paste it into a plain-text editor. Read the pasted version, not just the page on screen. Pay particular attention to job titles, certifications, skills, employer names, and words that seem oddly spelled. Check whether text is missing or appears in the wrong order.
-
Compare suspicious words with the visible PDF. An unexpected capital letter, punctuation mark, or missing letter may be easy to overlook. For instance, medicaIon can look like medication at a glance. Copying and pasting is a useful first check, but different software can extract text differently. A clean result does not guarantee that every employer's system will read the PDF correctly.
-
If you find errors, return to the original document rather than trying to fix the extracted text. Export a fresh PDF, and consider changing the export method or font if the problem persists. Then repeat the text-extraction check. Changing the font is a troubleshooting step, not a guaranteed fix.
-
Follow the employer's file instructions. If the application specifies PDF or DOCX, use an accepted format. When the system lets you review information it has filled in from your résumé, check those fields against the original before submitting.
A dedicated parsing diagnostic can provide another level of checking. Docent's ATS Diagnostic examines résumé readability and can identify certain text-extraction problems that are difficult to recognize through ordinary visual inspection. It does not reproduce every employer's ATS, but it can help identify issues worth correcting before submitting applications.
If a tool highlights unusual extracted characters or reports a possible parsing problem, investigate the underlying file rather than assuming the résumé contains spelling mistakes. The aim is not to make every document perfect for every possible parser. It is to avoid submitting a file with a known problem you had an opportunity to correct.
8. What this means for job seekers
A résumé should do more than look right. The words describing your experience and qualifications also need to survive the process of being read by software.
You cannot know in advance exactly how every employer's applicant tracking system will handle your file. But you can check for obvious extraction errors, take suspicious results seriously, and verify a corrected version before applying. That is a practical safeguard whether you're entering the workforce, changing careers, or applying for a senior role.
References and further reading
[1] Adobe / ISO, PDF 32000-1:2008, §9.10, “Extraction of Text Content.” https://opensource.adobe.com/dc-acrobat-sdk-docs/standards/pdfstandards/pdf/PDF32000_2008.pdf
[2] Apache FOP issue FOP-2701, documented Calibri/font ligature text-search behavior. https://www.mail-archive.com/fop-dev%40xmlgraphics.apache.org/msg27694.html
[3] Pega Support, Pega Platform 25.1.2 text copied from generated PDFs appears garbled or unreadable when pasted into external applications. https://support.pega.com/support-doc/pega-platform-2512-text-copied-from-generated-pdfs-appears-garbled-or-unreadable-when-pasted-into-external-applications
[4] Greenhouse Support, Unsuccessful resume parse. https://support.greenhouse.io/hc/en-us/articles/200989175-Unsuccessful-resume-parse
[5] Greenhouse Support, Why didn't the candidate's resume import correctly? https://support.greenhouse.io/hc/en-us/articles/200721684-Why-didn-t-the-candidate-s-resume-import-correctly