1A 95% ATS score measures the wrong thing well
An ATS score reports how many terms from the job description appear in your resume. That is the whole measurement. A 95% match means 95% of scanned terms were found. It carries no information about whether you can do the work, because it was never built to.
Keyword matching is a filter, not a signal. Its job is to remove the obviously irrelevant before a person opens the stack. Clearing it means you are in the pile. It does not mean you are in the conversation.
So a high ATS score with no interviews is not a contradiction, and it is not bad luck. It is a precise diagnosis: you cleared the layer that a machine runs, and you are losing at one of the two layers that people run. Those layers evaluate different things, and optimizing harder for the first one cannot fix the other two.
Three layers, three different questions
A resume must survive all three, and they do not want the same thing
~7 sec
a recruiter spends on the first pass through your resume
3 layers
between submission and an interview: parser, recruiter, hiring manager
1 of 3
is what an ATS score measures, and it is the least selective one
0
keyword-matching points awarded for evidence you did the work well
Optimizing the first layer past the point of clearing it produces no benefit. The remaining two are where applications actually die.
2What each layer is actually asking
The three reviewers of your resume have different jobs, different amounts of time, and different definitions of a good candidate. A resume tuned for one can fail the next.
The same resume, read three times
What passes at each layer, and what gets you cut
The parser
Seconds. Fully automated.
- Asks
Do the terms from this description appear in this document?
- Rewards
Present keywords, clean parseable structure, standard section headings.
- Cuts you for
Missing terminology, tables and text boxes that do not parse, graphics in the header.
- Cannot see
Whether any of it is true, or whether you were good at it.
Clearing this is table stakes, not an achievement.
The recruiter
About seven seconds. First human pass.
- Asks
Is this a legible, plausible career for this role?
- Rewards
Upward trajectory, recognizable or relevant employers, continuity, numbers in the bullets.
- Cuts you for
Obvious keyword stuffing, seniority two levels off, unexplained gaps, a format that hides the relevant line.
- Cannot see
The hard part of the job. They are matching a pattern, not evaluating the work.
This is where most high-ATS resumes die.
The hiring manager
Minutes. Reads properly.
- Asks
Has this person solved the specific problem I am hiring for?
- Rewards
A concrete system, a named constraint, a result with a number and a timeframe attached.
- Cuts you for
Responsibilities where outcomes should be. Listing Python instead of what you built with it.
- Knows
Exactly which part of this job is hard, which is why generic competence reads as no evidence.
Specificity is the only currency at this layer.
A resume can be optimized past the parser and be actively worse at the next two. Stuffed keywords are the clearest example.
3The four patterns that pass machines and fail people
Keyword stuffing
Managed project management of projects using project management methodologies scores beautifully and reads as gaming within one line. A recruiter recognizes it instantly, and it costs more credibility than the keyword was worth. Use the term once, in a sentence describing something you did.
Responsibilities instead of outcomes
Managed social media accounts and grew Instagram engagement 47% in six months contain the same keywords and carry completely different information. The first describes a job description. The second describes you. Only one survives a human read.
Right keywords, wrong level
A resume can match every term while showing scope two levels off in either direction. The parser has no concept of seniority. The recruiter reads titles and tenure first, and this is one of the fastest cuts they make.
Formatting that survives upload and not review
Tables, columns, text boxes, and header graphics can parse acceptably and still land in the reviewer's system as a scrambled column of text. Open your resume in a plain text editor. What you see there is closer to what they see than your PDF is.
Every one of these clears an ATS comfortably. Each fails within seconds of a person reading it.
4What to change, in the order that pays
The fixes are not equally valuable and most people do them in the wrong order, starting with keywords because that is the layer with a score attached.
Where the effort actually pays
Ranked by what it changes at the layer where you are losing
| What it fixes | Worth doing when | |
|---|---|---|
| Rewrite bullets as outcomesHighest return | Converts recruiter and hiring manager reads. A number and a timeframe on every bullet you can support. | Always. This is the single highest-value edit on almost every resume. |
| Check the seniority signalFast, often decisive | Stops the recruiter cut where scope reads two levels off the req. | When you clear ATS consistently and never reach a screen. |
| Plain-text the formattingCheap insurance | Guarantees the reviewer sees the same document you wrote. | Once, now, if your resume uses columns, tables, or a graphic header. |
| Add missing terminologyOnly to a point | Clears the parser for skills you genuinely have and simply did not name. | When a real skill is absent from the document. Never as volume. |
| Push ATS score from 85 to 95Lowest return | Nothing. You already cleared the filter at 85. | Effectively never. This is optimization theater with a progress bar. |
If you clear ATS filters consistently and still get no screens, more keywords is the one change guaranteed not to help.
5Why fit is a different measurement from keyword match
A keyword scan asks whether a word is present. A fit score asks how deeply the role needs a skill and how deeply you have it, which is a different question with a different answer for the same resume.
The engine reads the description and your profile together across five weighted dimensions, with Skills Match carrying 40 of the 100 points and assessing depth on a five-level scale rather than counting hits. It also weights the job's core defining skill two to three times heavier than the peripheral ones, so missing the headline requirement costs more than missing three nice-to-haves, which is exactly how a hiring manager reads it and exactly what a keyword count cannot represent.
The practical difference is what you can do with the output. A 95% ATS score tells you the terms are present. A fit score with a gap breakdown tells you which requirement is costing you 18 points and whether that gap is one you can close with a sentence or one that means the role is aimed at somebody else.
See the depth-based fit score and gap breakdown on a real posting.
Score a Job FreeWritten by
Jesse Johnson
Founder, ShouldApply
Founder of ShouldApply. I write about job search strategy, hiring, and how to spend your time on opportunities that actually fit. Full bio →
Keep Reading
Frequently Asked Questions
Both, but in the right order. Your resume needs to pass keyword filters to reach a human at all. Once it does, it needs to be compelling to a recruiter in 10 seconds and to a hiring manager in 2-3 minutes. The mistake is optimizing only for ATS, which produces keyword-stuffed, outcome-thin resumes that pass the filter and lose at human review.
Most companies with 50+ employees use some form of applicant tracking. Smaller companies often don't, which means your resume goes directly to a human. At startups and small companies, the human review criteria apply from the first glance.
Most ATS tools target 70-80% as the passing threshold. Above 80% and you're well past the filter. At that point, improving your ATS score further has no meaningful effect on interview rate. The investment is better spent on making the resume more compelling to a human.
Resume scanners predict ATS filter success, not interview success. They tell you whether your keywords match the JD. They don't tell you whether your outcomes are compelling, your experience level is right, or your formatting will survive the recruiter's 10-second scan. Use them for what they measure, not as a proxy for the full evaluation.
Standard reverse-chronological format, single column, clean section headers, no tables or text boxes, standard fonts. The safest file format is .docx or PDF with actual selectable text (not a scanned image). Most ATS systems handle both; some handle PDF inconsistently with complex formatting.
Free Tools
Related Posts
ShouldApply Outcome Data
Interview rates by fit score band from tracked applications, with the methodology and its limits.
Resume Keywords That Actually Matter
Which keywords signal competence to recruiters vs. which are just ATS filler.
Why You're Not Getting Interviews
A broader diagnosis of why applications aren't converting to screens.
How to Tailor Your Resume Without Starting Over
The 20-minute process for aligning a resume to a specific role.
Passing the ATS isn't the goal. Getting the interview is.
The fit score looks at Skills Match, Experience Level, and Seniority Alignment together, not just keyword presence. A more complete picture of where you stand.
Score a Job