Website Trust Scores Explained: How Ours Is Scored and What the Grades Mean

A website trust score is a single number that summarizes how much verifiable, public evidence supports trusting a website — the same way a credit score summarizes repayment risk. Ours runs 0-100, maps to letter grades from A+ to F, and is built from ten weighted dimensions, each graded separately with cited evidence. This post opens the hood completely: the dimensions, the exact weights, the grade bands, and how to read a report.

If you want to follow along with your own numbers, run a scan first — the report will make twice as much sense with your grades in front of you.

The credit score analogy, taken seriously

Credit scores work because they compress something complicated — a life of financial behavior — into a number that strangers can act on. No lender reads your whole history; they read the score, then the report behind it when they need detail.

Websites have the same problem in the other direction. Every buyer, partner, and donor has to answer "can I trust this site?" with limited time and no standard instrument. So they improvise: a glance at design, a skim of reviews, a gut call. A trust score formalizes that improvisation — same evidence a diligent stranger would check, but checked systematically, weighted explicitly, and cited.

The scale runs 0-100 and converts to grades the way you'd expect from school: 97+ is an A+, 93+ an A, 90+ an A-, 87+ a B+, 83+ a B, 80+ a B-, and so on down through the C and D bands to F below 60. In plain terms: A-grade sites are excellent, B-grade sites are strong, C-grade sites are average with visible gaps, and D-and-below sites have serious, specific problems the report will name.

The ten dimensions and their weights

The overall score is a weighted average of ten dimension scores. The weights are fixed, public, and sum to 100%:

• Business legitimacy — 18%. Is there a verifiable real business behind the site? Registration records, consistent contact details, named and findable people, a physical footprint that checks out.
• Online reputation — 15%. What do independent sources say? Review volume, recency, severity of complaints, owner responses, forum sentiment, and whether the pattern looks organic or manufactured.
• Visual design — 14%. Does the site look like something a real, resourced operation maintains? Hierarchy, consistency, imagery quality, polish.
• UX / conversion — 12%. Can a visitor actually accomplish things? Navigation clarity, mobile experience, form and checkout friction, load behavior.
• Transparency & disclosure — 10%. Pricing visibility, refund and privacy policies, terms, named ownership — whether the site volunteers what buyers need or makes them dig.
• Technical health — 8%. HTTPS and certificate quality, security headers, performance, broken links — the objective, measurable layer.
• Content quality — 8%. Specific, evidenced, expert writing versus thin generic filler.
• Social & press presence — 7%. Live profiles, third-party mentions, signs the operation exists beyond its own domain.
• Domain & company longevity — 5%. Domain age, archive history, consistency between the founding story and the paper trail.
• Financial signals — 3%. Funding, hiring, distress signals — mostly a "nothing weird here" check.

The ordering encodes a philosophy: who you are (legitimacy) and what others say (reputation) outweigh how you look (design, UX), which outweighs technical hygiene. A gorgeous site run by no one verifiable scores worse than a plain site with a real business and real reviews behind it — deliberately.

Each dimension gets its own 0-100 score and letter grade, so the report shows you not just where you stand but where the points are leaking. A site can carry an A in technical health and a D in transparency; the overall grade tells you less than the spread does.

Where the evidence comes from

Every scan researches the site's public record: the site itself, business registries, review platforms, search results, social profiles, archive history, and press. The critical design decision is that every dimension verdict cites its sources. If the report says your reputation grade suffered from unresolved complaints, it links to the complaints. If legitimacy dinged you for inconsistent contact details, it shows both versions.

This matters for two reasons. First, you can audit the audit — check any claim yourself rather than trusting a black box. Second, it makes the report actionable: a cited finding is a work order, while an uncited score is just a mood.

What the grades mean in practice

A range (90-100): the public record strongly corroborates the site. Typically established organizations with clean legitimacy, healthy review profiles, and no significant flags. A+ is rare by design.

B range (80-89): trustworthy with minor leaks — a thin social presence, a policy gap, middling performance. Most well-run small businesses land here, and the report will name exactly which dimensions are holding back the A.

C range (70-79): average, which on the modern web means "real but unconvincing." Usually multiple fixable gaps: sparse reviews, anonymous About page, hidden pricing, stale content. C-grade sites lose winnable customers daily without knowing it.

D range (60-69): serious credibility problems — the kind a cautious buyer walks away from. Often a compounding mix: no verifiable business identity and no reviews and technical neglect.

F (below 60): the public record contradicts trust or barely exists. Not necessarily fraud — brand-new sites with zero footprint can land here — but a stranger has no rational basis for confidence, and the report will say precisely why.

Two honest caveats. A high score isn't a guarantee of virtue — it means the public evidence supports trust, and evidence can lag reality. And a low score isn't an accusation — new and offline-successful businesses often score poorly simply because the web can't corroborate them yet. The score measures verifiability, and verifiability is fixable: our guide on improving your trust score covers the standard path.

Reading a report like an analyst

Read the overall grade last, not first. Start with the dimension spread and find your two heaviest-weighted weak dimensions — that's where the same effort buys the most points. Read the cited evidence for those two. Then sort the findings into what you can fix on-site this week versus what needs an off-site campaign (reviews, press, indexing) over a quarter.

And if the report needs to convince someone besides you — a client, a boss, a board — the grades and evidence are already structured for it. Our sister product StrategyPresentation turns a completed scan into a presentation deck for exactly that meeting: dimension grades as the diagnosis, cited findings as the proof, and the fix list as the plan.

A trust score doesn't create your credibility. It measures what strangers can already see — and gives you the itemized list of what they're seeing. The number is the summary; the report is the strategy.