How this book is built
Every claim cites a primary source.
This book compiles what named authorities — LinkedIn's own product documentation, the U.S. Bureau of Labor Statistics, LinkedIn's Talent Solutions and Recruiter teams, MIT News, Google, and CPRW-vetted job-search research — actually say about how a candidate becomes visible on a recruiter's screen. Every empirical claim in the book has a source entry with author, publisher, year, and URL. If a claim can't be sourced, it doesn't ship.
44%
of active job-seekers got zero interviews in the previous month.
Source: Jobscan 2025 · N=442
32%
have been searching for over six months.
Source: Jobscan 2025
26.6%
Tunisian graduate unemployment. 35.6% for women.
Source: Institut National de la Statistique · Q2 2026
2.2×
interview rate for candidates with an optimised LinkedIn profile.
Source: Jobscan 2025 · N=442
10.6×
interview rate for candidates whose headline names the target job title.
Source: Jobscan 2025 · N=442
+8%
more hires when algorithmic writing assistance is applied to the résumé. Wages +10%. Peer-reviewed RCT, N=480,948, online freelance marketplace — closest available proxy for a traditional-employer résumé effect.
Source: Wiles, Munyikwa & Horton · Management Science 2025
Independent effect sizes
Each move, in its own study, on its own metric. Effect sizes are per row, per study — they do not stack multiplicatively.
Source: Jobscan 2025 · Disclosed survey, N=442 · Interview rate
Source: Jobscan 2025 · Disclosed survey, N=442 · Interview rate
Source: Wiles et al., Management Science 2025 · Peer-reviewed RCT, N=480,948 · Hire rate
Source: AEJ: Applied Economics 2023 · RCT, N=2,021 firms · Callback rate
Bar widths are visual, per-row. The numbers on the right are the source of truth. Effects on different metrics (multipliers vs. percentage-point lifts) are not comparable across rows and do not multiply.
01
The evidence bar
Every chapter is written against a hard rule: number, publisher, year, URL. If any of the four is missing, the claim is cut.
Rule 01
Every stat is sourced.
Number, publisher, year, URL. If we can't source it, we cut it.
Rule 02
Primary sources first.
LinkedIn Economic Graph, LinkedIn Help Center, LinkedIn Recruiter Help, BLS, peer-reviewed studies. Not a blog citing a blog.
Rule 03
Rejected claims are named.
Popular career-advice statistics that can't be sourced (e.g. "60% of recruiter search weight is on the headline") are explicitly rejected in the per-chapter sources, not silently repeated.
02
What counts as a source
Two tiers are citable in the book. A third tier is documented and rejected.
-
LinkedIn's own product documentation, U.S. Bureau of Labor Statistics, peer-reviewed studies published in named journals, government regulator publications.
Load-bearing. Every headline claim in the book is Tier 1 wherever possible.
-
Nielsen Norman Group, Baymard Institute, Jobscan (CPRW-authored articles), MIT News summaries of published research, DataReportal, TGM Research, ILOSTAT country profiles.
Citable when a Tier 1 source doesn't exist yet, or when a Tier 2 practitioner has done the closest disclosed-methodology work.
-
Undisclosed-methodology blogs, marketing-agency posts without a study behind them, aggregator repeats of unsourced statistics.
Documented per chapter — the per-chapter sources file names each rejected claim so it's visible what didn't make the cut.
04
Sources by chapter
Each chapter's per-page sources file lists every claim, source, year, URL, tier, and notes on methodology. In the reader, footnote numbers on every claim link straight to the primary URL.
-
LinkedIn Help Center · LinkedIn Talent Solutions 2025 · LinkedIn Recruiter Help · Jobscan 2026
11 primary-source footnotes. See chapter 1 notes.
-
LinkedIn Help Center · LinkedIn Talent Solutions · Jobscan CPRW
34 primary-source footnotes. See chapter 2 notes.
-
LinkedIn (Tim Jurka, March 2026) · Rajkumar et al., Science, 2022 · AuthoredUp 2026
14 primary-source footnotes. See chapter 3 notes.
-
Greenhouse Support · Workday parser docs · BLS · Iowa State ATS study
39 primary-source footnotes. See chapter 4 notes.
-
Google's XYZ pattern · Jobscan CPRW · CareerBuilder audit
15 primary-source footnotes. See chapter 5 notes.
-
CIO Magazine certification-value survey · Coursera Learner Outcomes Report · Anthology Institutional research
12 primary-source footnotes. See chapter 6 notes.
-
MIT News · Rajkumar et al., Science, 2022 (N ≈ 20M) · Jobvite Recruiter Nation
19 primary-source footnotes. See chapter 7 notes.
-
No new empirical claims. Re-uses evidence from Chapters 1–7.
Explicitly framed as an execution calendar. See the "note on sources" at the top of chapter 8.
05
Landing anchors
The numbers used to anchor the price and the market context on the landing page.
-
Décret n° 2026-67, published JORT 30 April 2026 — 554.736 TND/month for a 48-hour week
Source for the price-anchor line on the landing ("$29 ≈ 91 TND — less than five days of Tunisia's minimum wage"). A three-stage hike runs to 611.520 TND by 2028.
-
Institut National de la Statistique (INS) Tunisia — 26.6% overall graduate unemployment, 35.6% for women graduates
Live INS dashboard, Q2 2026 quarter. Cite quarter + access date when using in permanent copy.
06
Versions & updates
The book is versioned. When a source updates — LinkedIn changes its guidance, BLS publishes a new labour report, a new peer-reviewed study lands — the affected chapter is revised, the version increments, and the original download link serves the new file.
Current version
Book v1.1 · Aug 2026
Every existing reader receives every future revision. The version stamp in the footer of every page names the current release. Substantive change history lives in the per-chapter sources files.
Questions about a specific source or claim? Email hi@thevisibilitystack.com — a real person replies within two business days.