How to Tailor a Resume to a Job Description in 20 Minutes

Most advice on this stops at "find the keywords and put them in," which is true, useless, and the reason people either burn an evening per application or give up and send the same file to everything.

Tailoring a resume means selecting and rewording experience you already have so it lines up with one posting's priorities. It is not writing a new resume, and it is not adding things you haven't done. Done properly it touches four parts of the document and leaves everything else alone — which is why it takes twenty minutes rather than an evening.

Below is the whole procedure, the part every other guide skips (what to do about a requirement you genuinely don't have), and a way to verify you actually hit the mark before you send it.

First you need a master resume

Never tailor from the last version you sent. Every pass deletes a bullet, and six applications later you are editing a resume that has quietly lost half your career. Keep one long file — every job, every bullet, every number, three pages, ugly — and cut a copy down for each application.

Building it costs an hour, once. Every tailoring pass after that is deleting and reordering instead of remembering.

• Everything goes in, including the jobs and projects you would normally cut for length. You cannot select from material you never wrote down.
• The numbers live here. Write down the real figures while you still remember them: headcount, budget, latency, churn, ticket volume, class size. Reconstructing them eighteen months later is how people end up guessing, and guessed numbers come apart in an interview.
• Keep two or three phrasings of your strongest bullets. The same piece of work reads differently to a team hiring for delivery speed and a team hiring for reliability.

Step 1: Find the five things the posting is actually screening for (5 minutes)

Read the posting twice. You are looking for repetition and position, not for a complete inventory of nouns.

1. Anything that appears in both the responsibilities and the requirements sections. Those two lists are often written at different times by different people. Whatever survived both is the real job.
2. The first two bullets of each section. The order is rarely random — it usually reflects what the hiring manager said first when someone asked them what they needed.
3. Named things: tools, methods, certifications, systems, regulations. Looker, HIPAA, Kubernetes, SOC 2, QBRs. These are the words a recruiter types into the search box.
4. Required versus preferred. "Preferred" is negotiable and should not consume any of your four edits.
5. Write the five down on one line. If you cannot get to five, the posting is generic and there is nothing here to aim at — see the last section.

Most of the words in a job description are not keywords

When I built the keyword extractor behind our free ATS checker, the thing that decided whether the output was useful was not the matching — it was the discard list. Telling someone they are missing Kubernetes is something they can act on. Telling them they are missing experience, strong, proven, collaborate, solutions or senior makes the whole report look broken, and one line of noise is enough to make a reader distrust the rest of it. Real keywords tend to have a shape: an all-caps acronym (SQL, AWS), a name with punctuation in it (Node.js, CI/CD, C++), a capitalized proper noun sitting mid-sentence. Adjectives and abstract nouns almost never qualify.

Step 2: Change four things, and leave the rest alone

This is the step that turns an hour into twenty minutes. Four places carry nearly all of the signal:

1. The summary line at the top. Two sentences, rewritten to name the thing this posting is about. Highest-leverage edit in the document and it takes three minutes.
2. The skills list. Same skills, reordered so the ones the posting named come first — and the ones it never mentions fall off the end.
3. The top three bullets of your most relevant role. Not all of them. Three. People read down a page and stop.
4. The order of items inside each section. Projects, skills, bullets within a role: put the matching one first. Reordering costs nothing and changes what gets read.

Everything below is off limits, and keeping it off limits is what makes the twenty minutes possible:

• Job titles, employers and dates. These get verified. Adjusting them is not tailoring.
• The numbers. A metric that shifts to suit the posting was never a metric.
• Anything you could not talk about for five minutes. If it survives onto the resume you have to defend it in the room.
• The lower half of the page — education, certifications, the roles from eight years ago. Almost nobody's decision turns on them, and that is exactly where tailoring time goes to die.

Step 3: Rewrite the top bullets in their words, not yours

The facts stay identical. What changes is which fact leads and which verb carries it.

A posting that repeats "onboarding" and "renewals"

❌ Before: Managed 40 enterprise accounts, including escalations, quarterly business reviews, renewals and onboarding for new logos.

✅ After: Onboarded 40 enterprise accounts and owned their renewals; ran quarterly business reviews and handled escalations.

Nothing was invented. Onboarding and renewals were already in that sentence — they were third and fourth in a list of four, sitting behind two things this particular posting never mentions.

The summary line, before and after reading the posting

❌ Before: Results-driven customer success professional with 6 years of experience building strong relationships and driving growth.

✅ After: Customer success manager, 6 years in enterprise SaaS — most of it onboarding and renewing accounts in the 200-to-2,000-employee range.

The first version fits every customer success posting written this year, which is another way of saying it fits none of them. The second is harder to write and much harder to skim past.

One prerequisite: if the bullets are weak before you start — responsibilities rather than outcomes — fix that first. Tailoring a vague bullet produces a vague bullet with better vocabulary. The rules for writing achievements cover that step, and there are annotated resume examples if you would rather work backwards from a finished one.

What to do about a requirement you genuinely don't have

Every guide on this topic skips the case people are actually in. You have three honest options and one bad one.

1. Name the adjacent thing, concretely. They asked for Looker and you have Tableau. Write the Tableau line specifically enough that a human reads "this person builds dashboards" — because that is the thing being screened for, and the tool itself is a week of learning.
2. Handle the gap somewhere a person is reading. A cover letter or the first five minutes of a call is where a missing requirement gets discussed. A resume is a filter; the gap does not belong on it.
3. Decide it is a real gap and move on. If four of their five are missing, a better-worded resume is not the thing standing between you and the job.
4. ❌ Listing it anyway. This is the one that gets people into rooms they cannot survive. Our own checker says it in one line under the keyword panel: only add what's true — keyword stuffing is easy to spot in an interview.

The posting asks for Looker; you have used Tableau

❌ Before: Proficient in Looker, Tableau, Power BI and SQL.

✅ After: SQL and Tableau — built and still maintain the weekly retention dashboard the sales team runs their pipeline reviews on.

The second version does not claim Looker. It makes the Looker requirement feel like a formality, which is the most you can honestly do with it.

Step 4: Check the coverage before you send

You have just spent twenty minutes editing on the basis of what you believe the posting emphasised. Re-reading your own resume is a bad way to verify that, because you already know what you meant by every line.

Put the posting and your resume text side by side and get the list of terms that appear in one and not the other. Treat coverage as a proofreading step, not a score.

Paste your resume text and the posting: you get keyword coverage, the specific terms you're missing, and whether a parser can even find your contact details and dates. No sign-up, nothing uploaded — it runs in your browser. — Run a free ATS check

What to do with the result: a missing term you genuinely have is an editing miss, so fix it. A missing term you do not have is information, not a task. And a high coverage number does not mean you are a strong candidate — it means the machine can tell what you are claiming. The reasons resumes get filtered are mostly parsing failures, not keyword failures, and no amount of tailoring fixes a resume the parser cannot read.

Using ChatGPT or Claude to tailor: one prompt, three failure modes

A large share of the searches around this topic now include "ChatGPT" or "prompt", and the honest answer is that a model is very good at one half of this job and dangerous at the other. It can read a posting faster and more evenly than you can. It cannot supply your facts.

The prompt that doesn't blow up

Paste the posting, then your resume, then this: Use only facts that appear in my resume. Do not add, infer or estimate anything. (1) List the requirements this posting repeats or puts first. (2) For each one, quote the line in my resume that already covers it. (3) List the requirements that nothing in my resume covers — do not write bullets for these. (4) Rewrite only the lines from step 2, keeping every number, date, title and employer exactly as written. Step 3 is what makes it worth running. It tells you where your real gaps are instead of papering over them, which is the opposite of what a model does when left to its own devices.

• It invents numbers. Hand it a bullet with no figure in it and you will get back "improved efficiency by 30%". Nobody measured that. Delete every number you did not personally supply.
• It edits things you did not ask it to touch. Titles get "clarified", dates get reformatted, an entire bullet quietly disappears. Compare the output against your master resume line by line rather than skimming it.
• It over-matches. Tell it to use the posting's language and it will use all of the posting's language — three bullets in a row starting with the same verb, the company's internal jargon showing up in a job you held four years ago. That reads as a template, which is the thing you were trying to avoid.

It is the same boundary as using AI on a cover letter: a capable editor and an unreliable witness.

When not to bother

Tailoring every application is a comfortable way to feel productive while applying to jobs you will not get. Three cases where the hour is better spent elsewhere:

• The posting is generic. No named tools, no specifics, responsibilities lifted from a job-family template. There is nothing to aim at — send the master-resume cut and move on.
• You came in through a referral. A person is already carrying your resume past the filter. Spend the time on what you will say to them.
• You are sending thirty applications a week. Then volume is the problem, not wording. Eight tailored applications beat thirty untailored ones and you get your evenings back.

And the honest limit on all of this: tailoring changes which of your real qualifications a reader sees first. It does not change whether you have them.

FAQ

How long should it take to tailor a resume to a job description?

About twenty minutes per application once a master resume exists, plus the one-time hour to build that file. If it is taking an hour every time, you are rewriting instead of selecting — or you are editing the last version you sent instead of starting from the master.

Is tailoring a CV to a job description different from tailoring a resume?

For the one-to-two page industry CV used in the UK, Ireland and much of Europe, no — same procedure, same four edits. An academic CV is a different document: it is a complete record rather than a pitch, so you tailor the cover letter and the research statement instead of pruning the publication list.

How many keywords from the job description should I include?

Do not count them. Counting is how people end up with a skills section they cannot defend. Cover the terms that are genuinely true of you, get each one into a bullet where it is doing work rather than sitting in a list, and stop there.

Can I use ChatGPT to tailor my resume to a job description?

Yes for reading the posting, for finding which of your existing bullets already covers each requirement, and for rewording those bullets. No for writing bullets about requirements you do not cover — that is exactly where it invents facts. Tell it explicitly not to add anything that is not already in your resume, then check the output for numbers you never gave it.

Does a tailored resume actually get past the ATS?

An applicant tracking system does not reject you. It parses your resume into a database that a recruiter then searches. Tailoring helps because it puts the terms they will search for into your document in a form the parser can read. It does not help at all if the parser cannot read the document in the first place, which is a separate and more common problem.

Originally published at resuhive.com.