Methodology

How the Awesomeness Score works.

No black box - the Awesomeness Score is a weighted audit across technology choice, version health, engineering quality and AI-readiness. Published, reproducible, and honest about its limits.

Four audited dimensions

One score, four dimensions. Three measure the engineering - technology, versions and quality; the fourth measures how ready the site is for AI agents. Each is measured on its own terms before they are weighted together.

01 · Technology

How good are the choices?

Every detected technology is ranked inside its own category by real signals - npm download volume and GitHub reputation. A popular, well-maintained framework scores high; a niche or abandoned one drags the category down. We measure the quality of the choices, not how many boxes you tick.

02 · Up-to-dateness

How current are the versions?

Each detected version is placed on its own release timeline. Running the latest release scores full marks; falling several versions behind costs points, because old versions are where known vulnerabilities and missing fixes live.

03 · Quality

How well is it built?

A Lighthouse audit - performance, accessibility, best practices and SEO count equally. PWA is deliberately out of the score. Where real-user data is available, performance reflects it. A fast page that fails accessibility still loses points, because the average drags it down.

How the weighting works

The four dimensions don't count equally - and neither do good and bad sub-scores. AwesomeTechStack weights weak results far more heavily than strong ones, because that is how real risk behaves. AI-Readiness joined as the fourth dimension in methodology 2026.08, under the same rules.

≤ 49×9Weak links dominate. A failing sub-score pulls hard.

50 - 88×3Middling results get middling weight.

≥ 89×1Strong scores barely move the needle - they are expected.

One outdated, vulnerable dependency hurts your score more than ten great choices help it. That is deliberate: the score should reward fixing the worst thing first.

Reading the score

0 - 49 low50 - 89 medium90 - 100 high

We publish this methodology so you can check our work. A dimension without a measurable basis - no resolvable versions, no rankable technologies, quality unavailable - drops out, and its weight renormalizes over the dimensions that remain; the report marks such audits as incomplete, because a score over three dimensions is not the same promise as a score over four. Scores move as the ecosystem moves: a version that is current today falls behind tomorrow, and every re-audit reflects that. We would rather show a hard number you can argue with than a vanity metric you can't.

Scoring v3-parity-2026.08 - rolling out; audits created before the rollout may still reflect the previous weighting.

04 · AI-Readiness

The fourth dimension: AI-Readiness

AI assistants and agents read the web differently. Every new audit earns an AI-Readiness grade, A-F, across six checks. It counts as the fourth dimension of the Awesomeness Score, under the same weak-link weighting as the other three - methodology 2026.08. As the AI web moves the grade re-checks, and the methodology version is announced right here.

llms.txt
Agent access
Agent identity
Structured data
Readable without JS
Sitemap

How we grade it →