# Verunel · decision brief

## Sample assessment: from measurements to next steps

Explicitly synthetic sample: three pages, four issue groups and four linked actions\. Start by investigating slow LCP and enforcing CSP; then distinguish the titles and verify the image purpose\.

- **Index:** 82 / 100
- **Usable coverage:** 95%
- **Issue groups:** 4
- **Cutoff:** 2026\-09\-07T12:00:00\.000Z

## Three priority decisions

1. **The main content appears too slowly** — Late main content can hinder the first view of the offer\. Business impact is not measured in this sample\. _development · M_
2. **CSP is only in report\-only mode** — Report\-only records violations but does not block disallowed sources\. This alone does not demonstrate an exploitable vulnerability\. _development · M_
3. **Two pages share the same title** — Distinct descriptive titles help people and search engines distinguish the content\. _marketing/SEO · S_

## What to approve now

- **Speed up the main content** — Late main content can hinder the first view of the offer\. Business impact is not measured in this sample\. _now · development · M_
- **Enforce the verified CSP policy** — Report\-only records violations but does not block disallowed sources\. This alone does not demonstrate an exploitable vulnerability\. _next · development · M_
- **Give pages distinct, descriptive titles** — Distinct descriptive titles help people and search engines distinguish the content\. _later · marketing/SEO · S_
- **Verify the image purpose and accessible name** — If decorative, the description unnecessarily repeats content in a screen reader\. No real user session was performed\. _later · design · S_

## Suggested ownership

- development: 2
- design: 1
- marketing/SEO: 1

## Limitations

- This is an authored demonstration, not a customer assessment\. Measurements and the displayed AI interpretation are synthetic\. It includes a limited selection of catalog inputs; remaining checks were not performed and the displayed coverage does not describe them\. Browser and LCP measurements cover only the entry URL; the additional pages are included in title comparison and HTTP/content checks\.
- LCP is a lab signal without element attribution or a causal timing trace\. Real\-user data and a comparison baseline are unavailable\. A technical score is not a compliance certificate\.
- The catalog contains 556 checks in 22 categories\. Scope/methodology and inventory explain execution; they have no score and do not affect the index\. The other 20 categories form 7 pillars\.
- An observation contributes: passed 1, warning 0\.5, failed 0\. A metric's normalized contribution is the mean of usable observations of the same metric and rule × 100, rounded to one decimal place\.
- Earned points = rule weight × normalized contribution / 100\. Pillar score = total earned / total maximum points × 100, rounded to an integer\. The index is the weighted mean of scored pillars, rounded again to an integer\. Rule weights come from the versioned presentation registry; weight 0 does not affect the score\.
- Unknown, not tested, blocked and error have no score outcome and reduce coverage; not applicable is excluded\. Metric coverage is usable / relevant observations, then weighted by rules and pillars\. The index is published only at 60% usable coverage or above\.
- Coverage describes the inputs to this particular assessment; it does not imply that every catalog check was performed\. Assessed URL count is a separate measure\. An absent baseline is neither improvement nor regression\.
- Status is a check outcome\. Contribution is a score calculation\. Severity describes the finding, priority orders recommended work, and coverage describes usable measurements\. These terms are distinct\.

---
Report: `VER-26-0085` · v4 · `2026-09-07T12:00:00.000Z` · `sha256:abb94f9d31b4129d502bd773568bfefdbc9cf56611733d2b7f23b74b400ff8b6`