Almost every hiring system offers to read your CV (or resume) and fill in the application for you. When it works it saves ten minutes. When it doesn’t, you spend those ten minutes untangling it, and then do the same on the next site, and the one after.
The good news is that most of the mistakes come from a handful of causes, and you can see them for yourself before any employer does.
What happens when you upload a CV
The software that reads your CV is called a parser. It does two things. First it pulls the text out of your file. Then it guesses what each piece of text is: your name, your phone number, a job title, an employer, a date, a school. It makes those guesses from headings, from the order of the lines, and from patterns such as “March 2021 – Present”.
It does not see your CV the way a person does. A person sees two neat columns. A parser sees text in a sequence, and if that sequence is jumbled, everything it builds from it is jumbled too.
First, a myth you can stop worrying about
You have probably read that “75% of CVs are rejected by software before a person sees them”. That figure traces back to a 2012 sales pitch from a CV-optimisation company that closed the following year, and no study behind it was ever published.
Hiring systems mostly store, sort and search applications. People decide. What a badly read CV really costs you is time, fixing fields on every form, and a messy profile in the employer’s database, where recruiters search for candidates by keyword long after you applied. Both are worth fixing. Neither is a secret robot rejecting you.
The two-minute test: see what the software sees
- Open your CV as a PDF.
- Select everything with Ctrl + A (Cmd + A on a Mac) and copy it.
- Paste it into a plain text editor: Notepad on Windows, or TextEdit on a Mac after choosing Format, then Make Plain Text.
- Read it from top to bottom.
This is a fair approximation of the first thing a parser does. If the text comes out in a sensible order, with your name, your contact details, and each job’s title, employer and dates sitting together, a parser has a good chance with it. If lines from two columns are interleaved, sections appear out of order, or something is missing altogether, that is what the parser gets too.
If nothing pastes at all, your PDF is a picture of text, usually a scan or an export from a design tool, and no parser can read it.
What confuses parsers, and what to do instead
- Two columns or a sidebar. The most common cause. Use one column, and move the sidebar’s contents, usually skills and contact details, into sections of their own.
- Tables and text boxes. Their contents are often read out of order, or skipped. Use plain paragraphs and bullet points.
- Contact details in the page header or footer. Some parsers ignore headers and footers entirely. Put your name, email and phone number at the top of the page itself.
- Creative headings. “Where I’ve made an impact” is lovely, but a parser is looking for Experience, Education and Skills. Use the plain words.
- Dates in mixed formats, or years only. Use one format for every job, with the month: “March 2021 – June 2024”, and “Present” for the job you have now. Application forms ask for months, so give the parser one to find.
- Each job laid out differently. Put the title, employer, location and dates at the top of every entry, in the same order each time. Consistency is what lets a parser tell where one job ends and the next begins.
- Icons, logos and skill bars. A parser cannot read a five-star rating for Excel, and most people can’t either. Say it in words: “Excel: advanced, including pivot tables”.
- The wrong kind of file. A Word document, or a PDF exported from Word or Google Docs, both work. A scan does not. If an employer asks for a particular file type, use that one.
You can keep a designed CV
None of this means giving up a CV you are proud of. Many designed CVs pass the test above, so check yours first. If it doesn’t pass, keep two versions: a plain, one-column version for applications, and the designed one for sending to people directly, or for attaching where a person, not a parser, will be the first to open it.
Check what the form filled, every time
Even a clean CV gets misread now and then. When a form fills itself from your CV, take two minutes before you move on:
- Is every job there, once, with the right dates?
- Is each description under the right job?
- Are your email address and phone number exactly right?
- Are your degree, school and dates in the right fields?
Fix it there. The employer sees what the form says, not what your file says.
How Draftloop handles this
Reading your CV again on every site is what makes the same mistakes come back. Draftloop reads it once, when you upload it, and turns it into your profile. You check that profile once and correct anything it got wrong. From then on, every application is filled from your corrected profile, so a mistake fixed once stays fixed.
When a form asks for your CV as a file, Draftloop attaches the original, byte for byte, so the person who opens it sees your layout exactly as you made it. And it attaches it to the CV field itself, not to the site’s own “fill in the form from my CV” box, which on some sites rewrites everything that has already been filled.