A RiskSignal Exposure Score comes from two things you can examine: a scan that measures the conversion path the way a shopper meets it, and a set of weights anchored to published research on what technical friction costs a store.
This page shows both: the research the weighting is built on, and the depth of the scan behind every finding. The weight values themselves stay internal, because they are the engine.
A surface checker loads a page once and reports what it sees. A RiskSignal scan is built to see what a shopper would actually hit on the way to buying. Open each stage.
The scan renders each store on an actual mobile viewport as well as desktop, rather than checking a desktop page and assuming the phone matches. When the two disagree, the worse view informs the read, because a store that works on desktop and breaks on mobile is broken where most of its buyers are.
Many stores load their real tracking only after a visitor accepts the cookie banner. The scan accepts consent and reads the page again, comparing what runs before and after. A check that stops at the banner never sees the tags that actually fire for a real shopper.
A script tag in the page source says nothing about whether tracking runs. The scan captures the network beacon a pixel sends when it records an event, so a finding of broken tracking means the beacon was watched for and never seen. That is the difference between a pixel that is present and a pixel that works.
Load timing is sampled repeatedly rather than trusted from a single reading. Bot-challenge and interstitial pages, the "Just a moment" screens that sit in front of some stores, are detected and thrown out rather than scored as if they were the storefront. A single sample against a challenge page would carry a meaningless number into the result.
Every weight in the score traces to published work on how technical friction affects buying. The four bodies of research below are the basis for how the score is weighted.
The number these weights produce is the Exposure Score, the primary, client-facing measure. REI (Revenue Exposure Index) is the scoring component inside Ægis, our diagnostic engine, that produces it. The weighting is deterministic: the same evidence produces the same score every time, from fixed weights rather than a black-box learned model.
Every score is a relative ranking against comparable stores. It does not promise a revenue figure or a conversion outcome.
The weighting is deterministic and disclosed in principle here. No black-box model sits behind the number.
Each finding reflects a single point-in-time scan. A store can change the day after it is measured.