You found the role on a Sunday evening, and it fit. The responsibilities read like a description of your last two years. You rewrote your summary, moved the most relevant bullet point to the top, uploaded the PDF and pressed submit. Three weeks later the portal still says "Received", and nobody has written.

After the twentieth application like that, or the fiftieth, the silence starts to sound like a verdict. Maybe your experience isn't as strong as you thought. Maybe you're too senior, or not senior enough, or you left that last job too soon. Most people who apply for a lot of jobs end up having this conversation with themselves, usually late at night, and they almost always cast themselves as the problem.

Some of those doubts deserve a fair hearing. Hiring is slow. Roles get filled internally, budgets get frozen, and plenty of good applications lose out to other good applications. But there is one cause people rarely consider, because you can't see it from where you sit: your CV may not be saying what you think it says.

Your CV is often read twice

When you apply through a job portal, the first reader is usually software. Many employers, and most large ones, use an applicant tracking system (ATS) to collect and sort applications. When you upload your PDF, the system pulls the text out of it and tries to file that text into fields: name, email, phone, job titles, employers, dates, skills. A recruiter searching or filtering candidates later is often looking at those fields, or at a plain-text copy of your CV, before they ever open the design you made.

The software never sees the page the way you do. It doesn't see the tidy sidebar or the confident heading. It sees a stream of characters, in whatever order it manages to get them out of the file, and it guesses what each piece means.

When the guess goes well, nobody notices. When it goes badly, nobody tells you.

What can go wrong inside a good-looking PDF

While building a CV app, we spent a lot of time pulling text out of PDFs with the same kinds of extractors that résumé parsers use. These problems came up again and again.

  • Columns read out of order. Two columns look orderly to a person. A parser that rebuilds the page from where each line sits can stitch the columns together line by line, so a sentence about your last role runs straight into a list of languages.
  • Skills filed under Experience. Parsers split the text at headings and file whatever follows under the last heading they saw. If your sidebar is read in the middle of your work history, your skills become part of a job.
  • A name or email that never comes through. Contact details tucked into a page header, drawn inside a graphic or set in an unusual way are sometimes skipped altogether. A recruiter can't call someone whose phone number isn't there.
  • Headings the parser doesn't recognise. "Where I've been" has charm, but "Experience" is the word the software is looking for. A heading it doesn't know can leave a whole section unlabelled.
  • Text that is really a picture. A CV exported as an image from a design tool, scanned on a printer, or saved with its fonts turned into outlines has no text in it at all. To a parser, the page is blank.
  • Dates and characters that get mangled. Some fonts draw a dash or a plus sign with a shape that has no character recorded behind it. The parser drops it, so a date range runs together into one long number and a phone number loses its country code.

None of this shows when you open the PDF yourself, which is what makes the problem so hard to spot: the CV that looks best to you can be the one that reads worst to a machine, and the only feedback you get is silence.

What the other side may have seen

If you have been blaming yourself, think about what actually reached the recruiter. Perhaps a profile with an empty name field. Perhaps your three years leading a team filed under Skills. Perhaps a summary that breaks off halfway through a list of software. Someone working through a long queue of applications won't stop to work out what went wrong.

Whatever they decided, they decided about a garbled copy of your CV that you never got to see. That doesn't explain every silence. It does mean the silence may say less about you than it feels like it does.

What you can do this week

The useful thing about this cause is that you can check for it yourself.

  1. Copy and paste your own CV. Open the PDF, select all, and paste it into a plain text editor. If nothing selects, your CV is an image. If the text comes out jumbled, a parser may see something similar. It is a rough test, since every parser reads a little differently, but it catches the worst problems in a minute.
  2. Use standard headings. Summary, Experience, Education, Skills. Keep the personality for the sentences underneath them.
  3. Put your name and contact details on the page itself, as plain text, rather than in a page header, a footer or an image.
  4. Write dates the same way every time, in a plain format such as "Mar 2021 to Jun 2024".
  5. Be careful with columns side by side. Some parsers rebuild the page from its layout and read straight across two columns, mixing them. One column reads right everywhere. If you love a two-column layout, check how it reads before you upload it.
  6. Look at what the portal filled in. Many portals fill in their own application form from your CV after you upload it. If your job titles or dates land in the wrong boxes there, that tells you something about how the system read the file.

None of this guarantees a reply. Good applications still go unanswered for reasons nobody controls. But it takes away one reason that has nothing to do with how good you are, and that is worth an evening.

Why we built a check for this

This is a large part of why we made PearlCV the way we did. Every PDF it makes has real, selectable text. Before you send a CV, it reads the designed PDF the two ways parsers read one: in the order the file stores the text, and rebuilt from the layout of the page. Then it runs seven checks on what comes out, from whether your name and contact details come through to whether each role and skill lands under the right heading and whether dates and characters read back intact. It shows you the text a parser actually gets, and tells you what to fix.

Every PearlCV design passes those checks as it ships, so one CV can go everywhere. If your own changes trip one, it says what and takes you to the fix, or makes an ATS-safe copy in the same design. The check, its fixes and the copy are part of PearlCV Pro; on the free plan, export shows you that the checks have run on your CV and says when something needs a fix. Everything happens on your iPhone, iPad or Mac, with no account to make. We wrote more about what hiring systems actually read.

If you have been applying for months and hearing nothing, spend ten minutes looking at your CV the way the software does. If it reads cleanly, you have ruled something out and can look elsewhere with a clearer head. If it doesn't, you may have found the reason some of those applications went quiet, and it is one you can fix.