How the recruiter heatmap works

The recruiter heatmap is the overlay we render on top of your CV when you toggle the eye icon in the editor. It paints warm blobs over the parts of the page our model predicts a hiring manager will fixate on during the first six seconds of a scan. This page is the full methodology. It explains what the model actually measures, what it doesn't, and how to read a heatmap honestly without over-trusting it.

What the heatmap is

The heatmap is a simulation, not measurement. We are not tracking a real recruiter's eyes. There is no camera. There is no live observer. What we do is apply a rule-based intensity model to your CV's DOM nodes, and draw the result as a warm-to-cool gradient overlay.

The reason a simulation is still useful: the eye-tracking research that grounds the model is reproducible and consistent across studies. The places recruiters actually fixate during their initial pass converge on a small, stable set of regions — and the variance between candidates is mostly explained by where on the page content lands, not by some recruiter-specific preference. A simulation that codifies those regions can give you a fast, repeatable read on whether your CV's first-fold composition is going to land or get skipped.

What it's grounded in

The intensity weights below come from a handful of widely-cited studies:

• The 2012 TheLadders eye-tracking study of 30 recruiters across 6 seconds per resume, which established the "F-pattern" attention map and named the six high-fixation zones (name, current title, current company, current dates, prior title, prior company).
• The 2018 Ladders update, which extended the original sample, confirmed the six-second average, and observed that polished, scannable resumes shifted recruiter time from 6 to 7.4 seconds — a meaningful gain when the bar is "did the recruiter pause."
• Indeed's 2021 hiring-manager survey, which corroborated the priority order (name → role → most recent employer → dates → education → skills cluster) using a different methodology (self-report rather than eye tracking).
• Recruiter coaching guidance from Harvard Business School, Stanford CDC, and Google Career Coach office hours, which converge on the same pattern.

We're transparent that these are not perfect proxies for every recruiter at every company. They're the best-validated structural patterns we have, and they're stable enough across studies that a rule-based model built on them is more useful than no model at all.

What the model actually measures

The heatmap source lives in src/lib/heatmap/model.ts in our codebase. It assigns each tagged region of your CV a baseline intensity:

| Region | Baseline intensity | |---|---| | Name | 0.95 | | Headline / title under the name | 0.85 | | Summary | 0.70 | | Experience section heading | 0.60 | | Most recent experience entry (role + company line) | 0.80 | | Older experience entries | 0.55 → 0.40 (tapering) | | Skills section | 0.55 | | Education heading | 0.55 | | Education entries | 0.50 | | Bullets (default) | 0.45 | | Sidebar lists (languages, certifications) | 0.35 |

Then we apply per-bullet adjustments. Bullets in your most recent role get +0.10. The first bullet in any role gets +0.10 more. And when your CV is paired with a job description in the workspace, every bullet's Wave 7b impact score folds in at 60% weight — so a bullet known to be strong glows hotter, a weak bullet cools down, and the heatmap stops being purely positional.

The final intensity, clamped to [0.15, 0.98], drives both the radius of the gradient blob (size of the hot region) and the colour (yellow at low intensity, orange in the middle, red at high). The lowest-intensity zones stay slightly visible — a CV that disappeared entirely outside the headline would be a misread, not a real signal.

How to read it honestly

A good heatmap has three properties:

1. The name and headline are unambiguously the brightest region. If they aren't, you have a layout problem — usually a header that's too crowded, a photo placed where it competes with the headline, or a name set at the same weight as everything else.
2. There's a clear secondary hot zone on your most recent role. This is the second fixation point in the six-second scan. If the page is uniformly warm, your most recent role isn't doing visual work, which usually means the role+company line is set too small or buried under a long entry above it.
3. Hot bullets cluster near the top of each role. Recruiters skim downward and stop reading after the first two or three bullets per role. If your strongest bullets — the ones with named systems, scale, and outcomes — are at the bottom of a role's bullet list, the heatmap will reveal it as cool zones above warm ones.

What the heatmap does not tell you:

• Whether the recruiter likes your CV. We model attention distribution; we don't model preference.
• Whether your ATS keyword coverage is right. Different layer entirely — that's the ATS Score Checker.
• Whether you got an interview. Layout that survives the six-second scan is necessary but not sufficient.

Why we built it before we built the methodology page

A reasonable reader may notice we shipped the heatmap toggle (Wave 7c) before this guide. That's deliberate. The model is small, transparent, and rule-based — anyone reading the source can verify the weights match what's documented here. We'd rather get the product into your hands and document the methodology as it stabilises than gate the feature on a perfect explainer.

If you find a case where the heatmap reads wrong — a recruiter who told you they fixated somewhere our model said was cool, a layout pattern we're missing — please send it to the product email. The model is data-grounded, and it should evolve with new data.