The '23 Trap: How Apostrophe Dates Bury Resumes in Workday

You wrote "Jan '20 to Mar '23" because it looked clean. Workday read it as noise, subtracted the role from your tenure total, and ranked you 400th out of 800 applicants for a job you can actually do. Nobody scrolled that far.

A Q1 2026 analysis of 10,000 ATS scans found 71.4% of resumes landed below the 75-point "qualified" threshold, and only 11.2% cleared 85. Formatting causes 23% of parsing failures, and date handling is one of the loudest offenders. This is the mechanical fix that moves the needle before you touch a single keyword.

Why apostrophe dates get you buried, not rejected

Apostrophe years like '23, season labels like Summer 2023, and year-only ranges like 2020 - 2023 cause Workday-class parsers to either drop the role from your work history or miscalculate its length, which then drags your total years of experience below the requisition's minimum. You are not auto-rejected. You are ranked too low for a recruiter working down the list to reach you.

The popular "75% of resumes get auto-rejected by ATS" line is folklore. The real mechanism is worse. Modern applicant tracking systems rank. Recruiters filter by a years-of-experience slider before opening a single PDF, and they work from the top down until the shortlist is full. If your parsed tenure comes back as 3 years when your actual tenure is 5, you never surface.

Here is what the parsers are actually doing under the hood:

• They regex on patterns like \d{1,2}/\d{4} (MM/YYYY) or Month YYYY.
• Anything that doesn't match one of the accepted patterns is skipped silently.
• Year-only ranges like 2020 - 2023 get treated as January 1 to January 1, understating tenure by up to 22 months per role.
• Season labels ("Summer 2023", "Fall '22") have no numeric month, so many parsers drop the role's dates entirely.
• Apostrophe years ('20 - '23) confuse the tokenizer, which sometimes reads them as possessives or contractions.

Santa Clara University's career center puts it plainly: "ATS systems often parse dates to calculate the duration of your employment. Inconsistent date formats can lead to errors in parsing, which might cause your experience to be miscalculated or ignored." That is the non-vendor version of the same warning Jobscan, HireFlow, and CareerScribeAI have all published against Workday specifically.

The Q1 2026 numbers on where resumes actually die

The median resume scores 62 out of 100 against its target job description, only 11.2% clear 85, and formatting accounts for 23% of parsing failures. Formatting is not the biggest problem. It is the cheapest one to fix.

The rest of the failure distribution matters because it tells you what order to work in. Missing keywords the candidate actually has, just phrased differently from the posting, drive most of the 77% that isn't formatting. So the date-format rewrite is a warm-up, not the whole fix. It gets you out of the 60 to 74 dead zone so the keyword work has something to build on.

If you are scoring in the mid-60s, a mechanical clean-up (dates, headers, contact info out of the page header, single-column layout) can plausibly add double-digit points. Resumeadapter's Q1 2026 pipeline data shows a median +17 point improvement after one cycle. That is enough to cross the qualified band on its own for a lot of candidates who are already close.

The supply/demand ratio that explains why nobody warns you

There are roughly 285 US software engineers competing on Workday tenure math for every one US professional who actually understands the platform's parser. That is why bad date-format advice persists and why the candidate is the only line of defense.

Refolk's index of professional profiles shows the imbalance clearly:

The top employers of that thin Workday bench in Refolk's index are Walmart (6 practitioners), PwC (4), Kainos, Capital One, and RGP. Those same companies run Workday tenants and receive your application through it. The person who could tell you exactly how your dates parsed is on the other side of the wall, and they are outnumbered 285 to 1 by the people whose resumes they are ranking.

The recruiter side is not much better. 99.7% of recruiters use an ATS to filter candidates, per Jobscan's survey of 384 recruiters. With 349,584 engineers chasing openings and 111,433 recruiters to look at them, no human is reading your PDF end to end. They are reading the parsed output. That output is only as good as your date formatting.

The five-minute date-format rewrite

Pick one date format and use it in every role, every education entry, every certification, and every project section. MM/YYYY (like 01/2023 to 02/2026) is the safest choice for Workday. Month YYYY (like January 2020 to March 2023) is a fine second option. Never mix.

Here is the rewrite, in order:

1. Replace every apostrophe year. '20 - '23 becomes 01/2020 to 03/2023. If you don't remember the exact months, use 01/2020 to 12/2023; overshooting the end month by a few weeks is fine, understating tenure by two years is not.
2. Expand every season label. Summer 2023 becomes 06/2023 to 08/2023. Fall '22 becomes 09/2022 to 12/2022. Parsers do not know that "Summer" is a month range.
3. Kill every year-only range. 2020 - 2023 gets treated as January 1 to January 1, costing you up to 22 months per role. Add months on both sides.
4. Standardize the separator. Pick one of to or - and use it everywhere. HireFlow's Workday guidance recommends spelling out to to sidestep dash-character parsing bugs.
5. Use "Present" for current roles, not "Current" or "Now". Workday recognizes Present. Other tokens sometimes get read as job titles.
6. Match the format inside your Education and Certifications sections too. A resume that uses 01/2020 to 03/2023 for jobs and 2019 for the degree still trips the mixed-formats flag that the atshiring.com Workday guide calls out for reducing timeline confidence.
7. Move contact info out of the header/footer while you have the file open. Parsers often skip page-header regions entirely, which means the perfectly formatted dates below don't matter if the parser can't find your email.

If you'd rather not do this by hand for every application, Refolk writes your resume from your own work history in the format Workday-class parsers expect, then rewrites it for each posting you apply to. Paste the job description, get back a version with the dates, keywords, and section structure that actually parses.

The knockout question layer sitting behind the parser

Even a perfectly parsed resume can fail a knockout question, which is a hard eligibility filter (years of experience, work authorization, location, degree requirement) that the employer sets on the requisition itself. Knockouts are the real gatekeeper, not the keyword scanner.

The way this compounds with bad dates is nasty. If a req sets a 5-years-experience knockout and your '20 - '23 role parses as roughly 3 years instead of the true 3 years and 4 months (plus your other roles), your total YOE drops below the threshold. The system flags you as unqualified before a recruiter opens the file. You get an automated "we've decided to move forward with other candidates" email the same day.

Two things follow:

• Answer the knockout questions on the application form honestly and completely, because those numbers override anything your resume implies.
• Make sure your resume dates add up to the same total years of experience you typed into the form. Discrepancies between the two get flagged for review, and "flagged for review" in a stack of 800 applicants means "not reviewed".

The Enhancv 2025 study of 25 recruiters across Workday, iCIMS, and Greenhouse put formatting at 23% of parsing failures. Content is the other 77%. So once your dates are clean, the next move is keyword alignment: pulling the exact phrasing from the posting into your bullets where your history genuinely supports it.

What "clean" looks like on the page

A Workday-safe work-history entry has the company, title, location, date range in MM/YYYY or Month YYYY, and bullets underneath, all in a single-column layout with no tables, text boxes, or graphics.

Compare the two versions of the same role:

The broken version is what a lot of designer-inspired templates ship with. The clean version is what the parser wants. There is no visual middle ground that keeps both happy, which is why the resume you use for LinkedIn's PDF export and the resume you upload to Workday should not always be the same file. Rewriting your history against a specific posting is the tailoring pass Refolk automates so you are not maintaining two files by hand.

FAQ

Does the date format really matter if I upload a DOCX instead of a PDF?

Format matters more with a DOCX, not less. Workday's parser reads DOCX structurally (paragraphs, runs, styles), and a stylized date like "Jan '20 - Mar '23" inside a text box or a table cell is even more likely to be skipped than the same string in a plain PDF paragraph. MM/YYYY in a plain paragraph is the safest combination regardless of file type.

Will my resume score jump if I only fix the dates?

Sometimes, and the mechanism is worth understanding. If your true years of experience was being understated by 1 to 2 years because of dropped or miscalculated roles, fixing the dates can move you across the requisition's YOE knockout in a single edit. If your dates were already close to correct, expect a smaller lift and focus next on keyword alignment, which drives most of the remaining 77% of parsing failures.

Is "Present" better than leaving the end date blank?

Yes. Blank end dates get interpreted inconsistently across Workday, iCIMS, and Greenhouse. Present is the token the major parsers explicitly recognize as "still employed here." Current and Now work in some tenants and get read as job-title fragments in others. Use Present and don't gamble.

How do I know if my current resume has this problem?

Open your file, search for an apostrophe, and search for any of the words Spring, Summer, Fall, Winter. If any of those turn up in your work-history dates, you have the problem. Then look for date ranges that are only years (2020 - 2023 with no months). Those are the three patterns that quietly cost you tenure, and the five-minute rewrite above fixes all three.