How Vance Rank finds storefront problems — and what it cannot see.

01What we inspect

A scan covers the storefront signals your customers and search engines actually encounter. Each rule is versioned, so “checked the same way” means the same rule over time — and a passing result is a real evaluation, not missing data.

02What makes a finding a finding

An observation becomes a finding only when it carries everything a decision needs:

• The observation — a stated condition in plain language, not a hint.
• The affected page — the exact URL where it was seen.
• The evidence — a captured artifact with an integrity hash and capture time, so the claim can be re-checked later.
• The severity — commercial, not aesthetic: Critical when a purchase path or whole-store visibility is blocked now; Major when discovery, trust or storefront operation degrades; Minor for hygiene.
• The timestamp — when it was true, because storefronts move.
The reproducible method — the same versioned rule can run again on demand. Verify exists precisely to prove “fixed” against your live pages.

The honest zero. “0 findings” means “not found in the pages assessed” — never “your store is healthy”. Pages that were not scanned cannot be claimed in either direction.

03Why there is no single “health score”

An arbitrary 84/100 tells you nothing about what to do on Monday morning. Scores average the important with the cosmetic: a storefront can score well while its checkout price markup silently contradicts the visible price — the kind of purchase-facing inconsistency a score hides and a Vance Rank finding names, with the page and the evidence attached.

Vance Rank deliberately reports concrete problems, their commercial weight, and their proof. If every item is closed, the health is real — not an average.

04From finding to resolution

Found

The scan captures the observation and its evidence.

Prove

Severity and scope are grounded in what the crawl actually observed — never inferred from a single page.

Fix

For eligible findings on a connected store, Vance Rank proposes the exact change — a diff you see before anything happens. Nothing is written without your approval of that exact proposal. Where a change is not supported, the finding still comes with the manual steps to close it.

Verify

After a change — yours, your developer’s, or an approved fix — the finding’s rule re-runs against your live pages and records before/after. “Done” means observed, not assumed.

Watch

Patrol rechecks the baseline on schedule so a quiet regression becomes a confirmed incident instead of a customer complaint.

05Patrol and Review, precisely

Patrol rechecks the recorded baseline on your plan’s schedule. Ordinary pages are examined the way a visitor’s browser sees them; robots and sitemap files over HTTP. A difference is not an alert storm: you are alerted only after a change is confirmed more than once, repeated detections of the same problem merge into one incident, and the incident closes only when a fresh observation proves the condition gone. You get one alert per incident opening and one when it resolves.

Review is records-only: a report built from up to 90 days of stored findings, verified corrections, incidents and daily status, published with explicit blind spots. It does not perform a new crawl — it reads what the watching already recorded.

06Scope and limitations

A public scan sees what a public crawler can see. To read results responsibly, know what stays outside its sight:

• Behind-login surfaces — cart internals, checkout pages, account areas — are not crawled.
• Variant-picker prices beyond the default selection are not enumerated.
• Highly interactive behaviour is checked through the raw and rendered page inspections described above — not simulated customer journeys.
• Anything deliberately hidden from a public crawler stays hidden.

07Scanning you can allow on purpose

Scans are read-only and need no store credentials. The crawler identifies itself as VanceRankBot/1.0 (+https://vancerank.com/bot), respects robots.txt, caps its request rate, response sizes and durations, and never attempts purchases. Quick Scan results stay held temporarily for up to 30 minutes and are never written to a store. See Security & data and Privacy.