What an applicant tracking system actually does with your résumé

An applicant tracking system is a database with a form on the front of it. That is the whole answer to what is an applicant tracking system, and almost everything frightening you have read about one is describing something it does not do.

It stores applications. It lets a recruiter search them, tag them, move them between stages, and send the rejection emails in a batch. Employers buy one for the same reason a warehouse buys shelving.

The fear attached to it comes from somewhere specific, and the source is worth knowing.

Where the number everyone quotes came from

You have seen the claim. Seventy-five percent of résumés are rejected by the software before a person ever sees them.

It traces to a sales pitch.

Two business academics, one at Mount Royal University and one at the Haskayne School of Business in Calgary, went looking for its origin and published what they found in The Conversation on 5 July 2026. Their sentence is worth quoting in full, because the whole edifice rests on it: "The statistic originated from a 2012 sales pitch by Preptel, a resume optimization company that went out of business the following year. No methodology was ever published, yet the figure has spread widely."

A company selling résumé optimisation produced the number that makes résumé optimisation necessary. Then it went under, and the number outlived it by more than a decade.

Now the part that should make you suspicious of us too. Go and search for that debunking and you will find dozens of pages repeating the Preptel story, almost all of them citing each other rather than anything Preptel published. We could not find a primary document either. So the honest position is narrower than the confident one: nobody has produced evidence for the 75% figure, and nobody has produced a primary source for the story about where it came from. Both of those are things people repeat because they read them somewhere.

That is the actual lesson, and it is bigger than one statistic. This category runs on numbers nobody checks.

What the software does do

Three things, roughly, and none of them is judgement.

It collects. The form you fill in writes a row. That is why the form asks you to retype your work history when your résumé is already attached: the résumé is a file, and a file is not searchable in the way a database column is.

It sorts. A recruiter can filter by anything the form captured. Years of experience, location, whether you answered yes to a question, when you applied.

It routes. Statuses, stages, automatic acknowledgement emails, and the rejection that arrives at 2am because someone cleared a queue before bed.

What it does not do is read your résumé and form an opinion. Some systems produce a match score, and a recruiter may sort by it. A sort is not a rejection. The application is still in the database, and whether anyone opens it is a decision a person makes about how far down the list to read.

What actually filters you

Something does filter you out automatically. It just isn't the PDF.

It is the questions on the form. Are you authorised to work in the United States, do you now or will you in the future require sponsorship, how many years of experience do you have with the one tool they care about. Employers configure rules on those answers, and the rule fires on what you typed into the box, not on anything in your document.

We watch this happen constantly. Our engine reads employer application forms all day, and the questions that decide an application are almost never about prose quality. They are yes/no items with a rule behind them, and a wrong answer to one of them ends the application regardless of how good the résumé was. Our own post on what actually gets a tech résumé read goes into the knockout question in more detail.

The other filter is arithmetic, and it is the one people underrate. When we counted 549,667 live US postings in August, 3.6% of software engineering roles were open to someone at entry level. Not 30%. If you are a new graduate applying to fifteen carefully chosen jobs, the thing limiting you is not a robot's opinion of your formatting.

So what should you actually do

Stop optimising for the software. It is not the audience.

Write the résumé for the person who will read it, which means specific claims with numbers attached to them rather than adjectives. Our guide to résumé verbs is the long version of that.

Then get the mechanical things right, because they are real and they are cheap:

• Answer every form question deliberately. That is where the automatic rules live.
• Submit a normal file. A PDF exported from a normal editor is fine. Text in an image is not, and neither is a two-column layout that scrambles when parsed.
• Do not put a percentage next to a skill. Nobody can defend the missing fifth.
• Apply to enough things that the entry-level ratio stops being the binding constraint.

And check that your applications actually landed. A form throws a validation error under the submit button, you had already switched tabs, and your tracker says applied because you told it you applied. That gap is the reason we built what we built: our free tier is 25 applications a month and nothing counts as done until there is a saved confirmation from the employer's own page.

The applicant tracking system was never the thing standing between you and the job. A person deciding how far down a long list to read is.