How Recruiters Actually Read a Resume: The Six-Second Scan
The first resume screen lasts seconds and covers six data points. What eye-tracking actually found, and how to put your evidence where it gets seen.
A recruiter’s first look at your resume is not a read. It is a triage decision, made in seconds, about whether your page earns a longer look. Your current title, employer, and dates do more work in that pass than any paragraph about yourself.
What the eye-tracking actually measured
The famous figure comes from the job site Ladders. In 2012, Ladders put eye trackers on 30 professional recruiters over a 10-week period and recorded how they reviewed resumes. Recruiters had estimated that they spent four to five minutes per resume. The trackers measured about six seconds for the initial fit or no-fit decision.
Ladders repeated the work in 2018 and reported an average initial screen of 7.4 seconds. That update also reported which pages held attention: simple layouts, clear headings, and bold job titles with bulleted accomplishments. The weakest pages had cluttered layouts, multiple columns, long sentences, and little white space.
Hold those numbers at the right size. Neither Ladders study was peer reviewed, the 2018 release did not disclose its sample size, and Ladders sells resume services. Setting matters too. In April 2018, StepStone ran eye tracking with 33 HR professionals in Austria viewing four sample CVs with no applicant queue behind them. Their mean viewing time was 43 seconds, against their own median estimate of two minutes.
The honest reading is not “you only ever get seven seconds.” It is that the first decision is fast, made from a small part of the page, and a page that passes it gets read more closely. For scale: a 2019 meta-analysis of 190 studies put adult silent reading of English non-fiction at 238 words per minute, so 7.4 seconds covers about 29 words. Screeners skim rather than read, so treat that as an order of magnitude: tens of words, not hundreds.
The six things that get the time
In the 2012 Ladders study, almost 80 percent of the initial review time went to six data points:
- Your name
- Your current title and company
- The start and end dates of your current role
- Your previous title and company
- The start and end dates of that previous role
- Your education
Beyond those fields, the report said recruiters did little more than scan for keywords matching the opening. That list explains most practical resume advice. Reverse chronological order is not a tradition. It puts the two roles screeners inspect at the top, in the order they inspect them. Consistent dates are not fussiness either: dates are two of the six fields, and a hurried reader should not hunt for them in a different place on every entry.
After the six fields, leftover attention goes mostly to the left edge of the page. Career advisers make the same point about bullets: put the important detail at the start of the line, because a skimming reader decides from the first few words whether the rest is worth finishing. A bullet opening with a duty says you held a job. One opening with a scope or a result says at what level you held it.
Volume has made this pass harsher. Greenhouse’s 2026 hiring benchmarks, built on more than 640 million applications across more than 6,000 companies, show applications per job rising from 116 in 2022 to 244 in 2025, while recruiting teams shrank from about 10.4 recruiters per organization to about 4.6. More pages per opening and fewer readers means more weight on the quick signals at the top of the page.
The software screen comes first, and it is dumber than advertised
Before a person scans your resume, software usually stores and parses it. This stage generates a lot of folklore, so separate what the evidence supports from what it does not.
A 2026 report from Huntr, based on 1.7 million applications, 225,000 resumes, and interviews with recruiters at Amazon, Microsoft, a Big Four firm, and a Fortune 500 manufacturer, found that no major applicant tracking system auto-rejects a resume without human review, and there is no universal score that follows you between companies. The main automated filter is blunter: knockout questions on hard requirements such as work authorization, a minimum degree, or a required certification. Fail one and your application can be set aside before a person opens your file. Keyword filters, in that report, were used mainly when recruiters search a database, not to screen a specific application.
Formatting still matters for a mechanical reason: a parser has to pull your titles, employers, dates, and skills into the right fields. The University of Virginia Career Center gives the standard guidance, and it matches the findings for human readers. Avoid images, columns, tables, and graphics. Keep key details out of headers and footers, where many systems cannot read them. Use standard headings such as Work Experience, Skills, and Education, simple bullets, left-aligned text, and a standard font at 10 to 12 points. Put relevant terms from the job description inside your bullets, where they describe real work, rather than only in a skills list. The same source notes that 97 percent of Fortune 500 companies use an applicant tracking system, so designing for the parser is the default case.
None of this rewards repeated keywords. In the Huntr data, tailored resumes converted to interviews at about twice the rate of generic ones. Tailoring means selecting and ordering your real experience so the match is easy to see, not claiming experience you do not have. A Robert Half survey reported in March 2026 found that 67 percent of HR leaders said the surge of AI-assisted applications has made it harder to verify skills. A claim that collapses under one interview question costs more than a modest claim that holds up.
Writing for the scan without gaming it
Start with the lines that get read, in this order.
Title, company, and date lines. Keep your official title exactly as your employer uses it, since verification can check it. If the title undersells the work, add context on the company line, such as the team or function, rather than inventing a grander title. Use one date format, in the same position, for every role.
The first bullet under your current role. Assume it may be the only bullet a screener finishes. Lead with the bullet you would keep if you could keep only one. As an illustration, not a real candidate: “Cut failed card charges 18 percent with a retry service for declined payments” gives a skimming reader a result and a skill in one line. “Responsible for improving the payment process” gives a duty and no evidence.
The opening words of every other bullet. Verb, then the thing it acted on, then the result where you have one. Drop filler openings that could fit any job anywhere.
The top of the page. If you use a summary, keep it to one or two specific lines naming your target role and your strongest relevant evidence. A paragraph of general traits is what a fast scan skips first, and it pushes your current title further down.
One column, plain structure. Single-column pages held attention in the Ladders work, give a reader one left edge to scan, and parse in the order you wrote them.
A seven-second test you can run tonight
Look at your resume for about seven seconds, taking in only your name, title, company and date lines, education line, and the first few words of each bullet. Then look away and write down what a stranger would conclude: what work this person does, at what level, and what they have done.
If that summary comes back as a job family and a list of duties, the fix is rarely a redesign. It is a short edit pass over a few dozen words: the header lines, the first bullet in each role, and the openings of the rest.
Bottom line
The first screen of your resume lasts seconds and runs on six fields: name, two titles with their employers and dates, and education. Design for that reality with an accurate title, consistent dates, a single-column layout a parser can read, and bullets that open with scope and results instead of duties. The numbers behind the scan come from small, older, non-peer-reviewed studies, so treat them as a design constraint, not a law. Earn the longer read first, then reward it. When the longer read becomes an offer, the negotiation runs in its own order; see how to negotiate a salary offer.