Most resume-tailoring advice treats keywords as a single set of nouns to copy from the job description. The reality is closer to two separate lists. One serves the ATS parser — the system that turns your resume into searchable metadata. The other serves the recruiter — the human who scans your bullets for credibility and fit. The lists overlap but aren't the same, and resumes that optimize for only one tend to underperform.

This post separates the two and shows how to satisfy both without doubling your workload.

What the parser is reading vs. what the recruiter is reading

The parser is a literal-match system. It indexes specific nouns — tool names, certifications, languages, role titles — and runs filter queries against them. "Snowflake" is in or out. "PMP" is in or out. If the posting requires a tool and your resume doesn't name it, you don't match.

The recruiter is reading for signal. They open the document for 7 seconds and look at the first bullet under your most recent role. They want to know what you actually did — and that means verbs and outcomes, not just nouns. A bullet that lists "Snowflake, Python, dbt" reads as a skills dump. A bullet that says "Built the analytics pipeline in Snowflake serving 80+ enterprise customers" reads as a person who did the work.

Two keyword sets, one resume

The same resume contains both. The trick is putting each in the right place:

The matched column — the parser keywords — should appear in the bullets that describe the work, in the skills section if you have one, and (sparingly) in the summary. Each tool or framework named on your resume should be defensible in a 60-second interview conversation.

The missing column — the recruiter keywords — is what most resumes underuse. Strong verbs and concrete scope are what differentiate your bullets from the candidate's whose bullets read like a job description. "Led", "shipped", "renegotiated", "owned" — each implies a specific kind of action the recruiter can picture.

Why the weights matter

A specific implication: optimizing only for parser keywords gets you past the ranking step but generates a resume that reads poorly when the recruiter opens it. Optimizing only for recruiter keywords means a strong-reading resume that may get filtered before a human sees it.

The 1:5 ratio is the constraint. Even if you nail the recruiter-side keywords, only ~20% of dashboard rows lead to the resume being fully opened. The parser keywords are what get you onto the right side of that ratio.

A practical workflow

Three places to think about each keyword set:

Layer 1 — Parser keywords (nouns):

1. Pull every tool, framework, certification, and standard label from the posting. Aim for 8–15 nouns.
2. Match each to a bullet on your resume where you'd reasonably mention it. Where the noun applies and you've done the work, name it the same way the posting names it.
3. Verify the skills section reflects the same nouns. Don't dump all 15 — pick 8–10 you can defend in a 60-second conversation.

Layer 2 — Recruiter keywords (verbs + outcomes):

1. Read your most-recent role's bullets out loud. Strike any that start with "Responsible for," "Worked on," "Helped with."
2. Rewrite the openers with concrete verbs — "shipped," "led," "renegotiated," "cut," "built."
3. Add scope or outcome to each: a number, a delta, a count, a date, a downstream effect.

Layer 3 — Skills-section discipline: A skills section that lists 40 items reads as either spam or undifferentiated to recruiters. 12 well-chosen skills, each defensible, reads as a person who knows their work — and still gives the parser everything it needs.

What this isn't

A few clarifications:

• It's not an argument for stuffing. Both keyword sets only work when the underlying work is real. See keyword-stuffing-vs-keyword-fit for the line.
• It's not "ignore one layer." Parser keywords get you into the room; recruiter keywords get you out of it with an interview. Both matter.
• It's not a different process for different industries. Sales, engineering, design — the same two-layer structure applies. The nouns and verbs change by role, but the layering doesn't.

The summary: a resume that scores well in both layers — the parser's literal noun-match and the recruiter's verb-and-outcome scan — outperforms one that wins at only one. Most tailoring advice covers Layer 1 well. Layer 2 is where the underused leverage is.