Website and Content-Tree Compliance Scanning: Static, Dynamic and Multilingual Review

Compliance follows the material, not the file format

A customer claim can appear in a film, subtitle, PDF, campaign image, CMS fragment or web page. The underlying compliance question is similar: what is present, where does it appear, which rule applies and who must decide the action?

For video, the evidence location is usually a timecode, frame or subtitle cue. For a website or connected content system, the location may be a URL, route, page section, DOM element, CMS field, document paragraph, image region or rendered screenshot.

The evidence model remains consistent:

Start with an authorised content tree

A useful scan begins with scope, not an unrestricted crawl. The content tree is the inventory of pages, routes, locale variants, components and linked materials that the customer has authorised for review.

Discovery inputs can include:

• One or more approved domains and starting URLs.
• XML sitemaps, route manifests or application inventories.
• CMS or Adobe Experience Manager exports and content-fragment dictionaries.
• Approved API feeds, repositories or structured metadata.
• Locale mappings and the designated source language or canonical record.
• Defined user journeys for content that appears only after interaction.
• Explicit exclusions, rate limits, authentication boundaries and retention rules.

Each discovered unit can retain its parent-child relationship. That makes it possible to review a whole branch—such as a market, campaign, product family or language—without losing the source page or component.

Static crawling covers pages and linked materials

A static crawl follows the authorised link graph and reviews the content returned for each route. This is appropriate for public pages, help centres, campaign sites, product information, downloadable PDFs and other materials available without interactive state.

The scan can capture:

• Page title, headings, visible copy and structured metadata.
• Images, alternative text, captions and linked documents.
• Canonical and language relationships.
• Broken or conflicting references inside the authorised scope.
• Exact passages or elements that match a configured compliance or content-quality rule.

Static crawling is repeatable and efficient, but it cannot see every state of a modern application.

Dynamic review follows defined journeys

Client-rendered applications can change after JavaScript runs, a filter is selected, a location is entered or an account state is applied. A meaningful review may therefore need a defined synthetic journey rather than a simple URL request.

A journey can describe authorised steps such as:

1. Open a specified route in a named environment.
2. Select a market, language, product or audience state.
3. Use search, filters, tabs, menus or pagination.
4. Submit approved test data to a form.
5. Capture the rendered page, accessible text, DOM and screenshot evidence.
6. Record the state and steps required to reproduce the finding.

Authenticated journeys can also be reviewed when credentials, OAuth and permitted routes are explicitly configured. Vidcomply does not bypass access controls; private-source integration is part of authorised onboarding.

Review connected materials as one evidence graph

Website copy is often assembled from several sources: a CMS fragment, product database, translation memory, image, document, campaign feed or metadata dictionary. Reviewing only the final page can identify a problem but may not reveal where it originated.

Vidcomply can normalise findings across configured sources so teams can distinguish between:

• A page-level issue created by presentation or rendering.
• A reusable CMS field affecting many pages.
• A document or image carrying outdated information.
• A metadata value conflicting with the visible claim.
• A translation difference affecting only one locale.

This turns a collection of pages into a reviewable content tree: the team can correct the source once, identify affected descendants and re-check the relevant branch.

Find cross-language discrepancies without treating translation as word matching

Language review is not limited to spelling or literal translation. Two pages can use different words and still communicate the same approved meaning; they can also look fluent while disagreeing on a critical claim.

When the source and target records are aligned, configurable checks can identify:

• Missing paragraphs, warnings, conditions or legal qualifiers.
• Conflicting prices, dates, measurements, product names or eligibility statements.
• Untranslated source text or mixed-language fragments.
• Obsolete terminology, disallowed wording or inconsistent brand names.
• Links or calls to action pointing to the wrong market.
• Images or documents that do not match the selected locale.

The output should show the source passage, target passage, location, difference and applicable rule. A language or policy owner then confirms whether the variation is intentional.

Detect outdated information using an explicit source of truth

No system can determine that information is outdated without a reference. Teams first define authoritative records, validity dates, approved terminology or conditions that make a claim time-sensitive.

Examples include:

• A campaign end date or expired offer still visible on one locale.
• A former product name remaining in PDFs or image copy.
• An old policy statement conflicting with the current approved version.
• A route, service, office or partner reference that has changed.
• Different contact details across pages, metadata and downloadable documents.

Vidcomply can flag the conflict, identify every affected location and preserve the comparison evidence. The authorised content owner decides whether to update, remove, exempt or investigate the material.

What a website compliance finding should contain

Results can be filtered by domain, branch, locale, content owner, rule, severity or status, then returned through reports or structured integration. The purpose is not simply to count defects; it is to route each issue to the team that can verify and correct it.

Frequently asked questions

Methodology and scope

• Authorised scopePublic crawls, private sources and dynamic journeys are limited to customer-approved domains, routes, credentials, environments and retention settings.
• Evidence modelVidcomply separates observed material, configured rules and final human decisions. A detected discrepancy is a review signal, not an automatic legal conclusion.
• Source of truthCross-language and outdated-information review requires identified source records, effective dates, terminology or comparison rules supplied or approved by the customer.