How the score is built

A candidate is rejected when it is:

• The company itself, or a former name, alternate name, or redirect of it.
• A parent company or a subsidiary of it.
• A marketplace, directory, or review site (for example G2, Capterra, Crunchbase).
• A social network, video platform, or news outlet.
• A bare conglomerate homepage (for example microsoft.com) — a specific product site under it (for example dynamics.microsoft.com) is a valid competitor.

Radar

Six axes, one per pillar. Your polygon (aqua) and every rival in your set (indigo through magenta, ranked closest first) are overlaid on the same six axes, so the shape itself shows where you lead or trail at a glance.

Leaderboard

Every company in your set, ranked by score, one bar per row — you and every rival found. Each row shows its rank, score, band, and the points gap to you. A pillar section repeats this for that one pillar’s scores instead of the overall total.

Profile line

The same six pillars laid out left to right instead of around a circle, every company in your set as its own line. The gap between your line and your closest rival’s is shaded in the lead or trail colour for that stretch — a wider shaded band is a wider gap; every other rival draws as an unshaded reference line. The same chart can switch to a single pillar’s individual checks.

Trend lines

One line per company in your set, overlaid: how often each homepage changed, one point per year for the last eight years, and Publishing days per month, one point per calendar month for the last twenty-four months — the count of days in that month with at least one dated sitemap or feed entry, not a count of posts. A year whose true count could be higher than shown (the public archive’s own row cap was hit) carries a ≥ marker instead of a plain number. A month before the observation horizon, or with no dated entries at all, is shown as unknown, not zero.

Sector bars

Inside each pillar, one bar per check per company in your set, sized by the points earned as a percentage of that check’s own maximum (not the raw point gap, so a small check and a large one are compared fairly). Each bar is coloured by that company’s own series colour — the same colour that names it in the leaderboard and every other chart — never by a lead/trail/even verdict.

Web Experience

Deterministic checks on the live site plus Google PageSpeed Insights (mobile).

When PageSpeed is unavailable, the 40 non-PageSpeed points are renormalised to 100 and confidence drops to 0.4.

Findability

Deterministic checks on the live site plus the PageSpeed SEO category.

When PageSpeed is unavailable, the remaining 75 points are renormalised and confidence drops to 0.7.

Content & Thought Leadership

Deterministic checks on the live site — sitemap, feed, and page structure.

Confidence is 0.9 when a sitemap or feed was parsed, 0.6 when cadence is read from HTML dates only.

Social Presence

Profile links on the site plus researched audience and posting cadence.

Assessed only when the social research step succeeds. A successful search that finds no accounts is a legitimately low score, not a missing pillar — every researched number carries its citation URL, and a number without one is dropped.

Paid & Press

A newsroom check on the site plus researched press coverage and paid-ad evidence.

Assessed only when the paid & press research step succeeds.

MarTech & Data

A curated, MIT-clean signature list of roughly 60 marketing tools, matched against script sources, inline globals, cookies, headers, and meta tags found in the page.

Confidence 0.85. Known limit: only tags present in the static page source — script URLs, inline globals, cookies, headers — are visible. Tags injected at runtime by a tag manager can be missed; the report states "detected in page source."