The Resume Standard: What a Good One Has to Contain

This guide gives you a gradeable standard for a resume: a set of criteria specific enough that two people scoring the same document land in the same place. It is for anyone writing or rewriting a resume who is tired of contradictory advice and wants a definition of done they can adopt as policy. Read it once to grade what you have, then keep it open while you fix the lines that fail.

Most resume advice argues about taste. A standard does not. It states what a resume must contain, what each criterion proves, and what it looks like when a criterion is satisfied on the surface but lies underneath. The evidence below is uneven in quality, and I flag which parts are solid and which are folklore, because a standard that overclaims is worse than none.

What a "good" resume must contain

A good resume is reverse-chronological, headed by role blocks that each carry employer, location, title, and dates, with three to five bullets per role that each begin with an action verb and quantify an outcome. That is the load-bearing consensus across primary sources, from the Harvard College resume guidance to the wider resume-writing field.

The most useful thing about this consensus is that it is checkable. Each criterion below can be scored yes or no by two different readers without a conversation about style.

• Order: reverse-chronological, most recent role first.
• Role header: employer, location, title, and dates on one line, parseable at a glance.
• Bullets: three to five per role, phrase form, no personal pronouns, starting with an action verb.
• Quantification: each recent bullet names a number or a concrete result.
• Length: justified by years of relevant experience, not padded.
• Layout: single column, white space, standard fonts at 10 to 12 point.
• Machine-readable: clean text extraction, standard headings, terms that mirror the posting.

The one criterion that people fight over is section order. Harvard and academic templates put Education first because for a student it is the strongest evidence. General-market sources put Experience first and Education last for working professionals. Both are right for their reader. The rule that resolves it: lead with your strongest credential. For most professionals that is experience; for students and recent bootcamp graduates it may be education or projects.

The four-part bullet, and how to tell a duty from an achievement

Every bullet follows the same four-part structure: action verb, what you did, a quantifiable result, and why it mattered. This is the criterion that separates a resume that reads like a job description from one that reads like a record of impact.

The failure is predictable. Most applicants get the first two parts right and skip the last two, which is why their resumes read like duties instead of accomplishments. "Managed the onboarding process" has a verb and an object. It has no result and no reason it mattered. It is a duty wearing a verb.

Here is what each part proves, and what it looks like when it lies:

• Action verb proves you did the thing rather than watched it. It lies when it is a soft verb like "helped" or "assisted" that hides who actually owned the work.
• What you did proves scope. It lies when it is so vague ("various tasks", "cross-functional initiatives") that no reader can picture the work.
• Quantifiable result proves the work moved a number. It lies when the number is invented; a metric that collapses under an interview question is worse than no metric.
• Why it mattered proves the number was worth moving. It lies when it is bolted on with an empty phrase like "driving business value".

[Action verb] [what you did] [quantified result], [why it mattered].

Rebuilt the weekly reporting pipeline, cutting report turnaround from 3 days to 4 hours and freeing the analyst team for forecasting work.
Negotiated three vendor contracts, reducing annual tooling spend by 18% with no loss of coverage.

The largest fixable edge on the whole page is quantification, not design. In one analysis, only about 26% of real resumes included at least five measurable results. That means a candidate who quantifies every recent bullet clears a bar most competitors leave open. Vendor sources put the uplift as high as 2.5 times more interview invitations, attributed to a LinkedIn Talent Report, but a precise, rigorously measured multiplier is not publicly established, so treat those numbers as estimates. The direction is safe; the exact figure is not.

The seven-second scan, and why it does not mean one page

The first gate is a scan of a few seconds, and it fixates on a fixed set of points: name, title, company, dates, previous role, and education. Because those points are fixed, "did the top third show them" is a criterion two reviewers can score identically.

A 2018 eye-tracking study measured an average initial screen of 7.4 seconds, up from about 6 seconds in 2012. Treat the exact number with care: the sample was only 30 recruiters, and how they were selected is unknown. The durable finding is not the stopwatch reading; it is which points draw attention. Roughly 80% of initial attention lands on those six fixation points.

Here is the trap. Writers assume a seven-second glance demands one page, so they cut. But the scan and the read are two different gates. Once a resume passes the glance, it gets a real read. In the hiring simulation behind the length data, reviewers spent roughly 2 minutes 24 seconds on one-page resumes and about 4 minutes 5 seconds on two-page resumes. The glance rewards a clean top third. The read rewards substance. Optimising only for the glance produces a thin resume that dies in the deeper read.

How long the resume should be, banded by experience

Length is set by years of relevant experience, not by a one-page rule. The one-page rule is a myth built on the scan number, and the largest empirical study points the other way.

In a hiring simulation, reviewers chose from 7,712 resumes, of which 5,375 were two pages, meaning recruiters preferred two-page resumes 2.3 times as often as one-page ones. More useful than the headline is that the preference scales with seniority. The same criterion cannot be applied flat: a one-page VP resume can read as a negative signal, while a two-page new-grad resume can read as padding.

Column source: ResumeGo hiring simulation.

Translate the gradient into a rule you can grade:

• Under three years of experience: one page.
• Three to ten years: whatever fits without cramping.
• Over ten years, or managerial: two pages by default.

The ATS reality, and what to do about it

An applicant tracking system was detected for 97.8% of Fortune 500 companies, so at large firms you should assume one is in the path. But it rarely auto-rejects. Contrary to popular belief, an ATS mostly acts as a search engine that lets recruiters filter and rank applicants by keywords and knockout requirements.

There is a widely repeated claim that about 75% of resumes are rejected by an ATS before a human sees them, sometimes attributed to Harvard Business School. Treat it as contested. It is repeated across vendor pages without a rigorous public source, and it contradicts the more careful description of how these systems are actually used. The real pressure is volume: a corporate posting draws around 250 applicants on average, so the system's job is to rank a large pile, not to slam a door.

That changes what you do. The mechanism rewards a resume that parses to clean text and mirrors the real terms of the specific posting, not one stuffed with keywords to beat a filter. Two rules follow:

• Parse clean. Use standard section headings, avoid tables and graphics that break text extraction, and keep a single column, which works everywhere and does not shrink on phones.
• Mirror the posting. Use the posting's own words for skills and tools where they are true of you, so the ranking picks you up for the right query.

Tailoring the same resume to every posting's real terms is slow by hand, which is where I can help. Refolk writes your resume from your own history, tailors it to each posting you apply to, and scores how well you actually fit the role, so the machine-readable and keyword-mirror steps happen without a manual rewrite each time.

The two-reviewer grading procedure

Grade a resume by scoring it against fixed criteria with two independent reviewers, then reconcile. The standard works because the criteria are concrete enough that disagreement points to an ambiguous criterion, which you then tighten. Follow the steps below in order; the first four build the resume, the last four test it.

The reconciliation step is the point of the whole thing. When two reviewers score the same resume and disagree, the disagreement is data. Either one reviewer misread, or the criterion was ambiguous. If it was ambiguous, rewrite the criterion until it is not. That is how a personal habit becomes a policy a team can adopt.

How this goes wrong: the failure modes

The most valuable part of a standard is the list of ways a resume passes on the surface and fails underneath. Each item below is a false positive: something that looks like it meets the standard but does not. For each, I give the check that catches it.

Chasing the seven-second number

A resume "optimised for the scan" can be thin on substance and die in the deeper read. Check: run the page-two test on every line - would a recruiter reviewing your application for this specific role consider it relevant and valuable? If not, it is decoration.

Believing the ATS auto-rejects

This leads to keyword-stuffing to beat a filter that mainly ranks and searches. Check: does the resume parse to clean text, and do its terms mirror the posting's real language rather than a keyword dump?

Two pages padded with filler

Length gets read as depth. Check: could you delete a full third of the content and lose nothing a recruiter needs? If yes, you are padding.

Fake or unverifiable numbers

An impressive metric that collapses in an interview or reference check does more harm than a modest true one. Check: can you back each figure with data or a reference? If not, remove or soften it.

Duties dressed as achievements

The verbs are present but the result and the reason are missing. Check: does each bullet name an outcome and why it mattered, not just an activity?

Two-column or graphic layouts

They look polished and break parsing and mobile rendering. Check: single-column layouts work everywhere; if the resume uses columns, sidebars, or text boxes, rebuild it in one column.

Trusting old LinkedIn multipliers as fact

Figures like "profiles with a photo get 21 times the views" are old, LinkedIn-attributed, and hard to pin to a live source. Check: label such figures as legacy or vendor folklore, not evidence. The direction (a complete profile helps sourcing) holds; the specific multiple does not.

Who is actually reading it, and syncing the profile layer

Understand who grades your resume before you decide how to write it. Most resumes are judged by hiring-side readers working under volume pressure, not by coaching-side experts with time to appreciate nuance.

The supply is lopsided. In Refolk's index of professional profiles, US recruiters and technical recruiters vastly outnumber the resume writers and career coaches who help candidates prepare.

Column source: counts from Refolk's index; the ratio is derived from those counts.

The takeaway from the 21-to-1 ratio: write for a fast, hiring-side reader under load, not for a coach with a red pen. That is why the scan matters, why the top third has to do work, and why a single true number beats three vague claims.

The profile layer sits alongside the resume for sourcing. Be skeptical of the exact multipliers attached to profile photos and completeness; those figures are old and LinkedIn-attributed. The direction holds: a complete, aligned profile helps recruiters find you. The gradeable criterion is consistency, not decoration - your public headline, current title, dates, and top experience should not contradict your resume. If a recruiter finds you through your profile and then reads a resume that tells a different story, that gap is the problem, not the photo.

If you want to see the hiring-side readers your resume will face, run this in Refolk:

The grading checklist and how to keep it current

Use this checklist as the two-reviewer instrument. Score each item yes or no. Any "no" is a fix; any disagreement between reviewers is an ambiguous criterion to tighten.

Two parts of this standard are time-sensitive, so re-check the mechanism rather than trusting a stored number. First, the scan-time and length findings rest on studies with real limits: the seven-second figure came from a 30-recruiter sample, and the length preference from a single hiring simulation. If you see a newer, larger study, weight it above these. Second, ATS behavior evolves, and market-share leaders shift; verify that your file still extracts to clean text by pasting it into a plain-text editor and checking that headings, dates, and bullets survive.

What to do next is simple. Grade what you have against the checklist today, fix every "no", then re-grade with a second reader before you send. Keep the four-part bullet and the experience-banded length rule as your defaults, and re-run the parse test each time you change the layout. The standard does not expire; only the studies underneath it will, and you now know how to check them.