Site types
Auditing a marketplace or listing website
Marketplaces, directories, job boards and property portals share one rule module, because they share one shape: many listing pages, each describing an item someone else offers. The audit checks that listing pages carry structured data, name who is offering, give a way to contact them or apply, and that the site offers search or filters and pagination. A marketplace is also weighted as a selling site in the score; directories, job boards and property sites are not.
Which sites get the listing rules
Four detected categories run the same module: marketplace, directory, jobs and real estate. The category engine recognises them from their own signals — seller and vendor paths such as `/sellers` or `/vendors` and wording such as "become a seller" for a marketplace; `/directory` or `/annuaire` for a directory; `JobPosting` markup and paths such as `/jobs` or `/careers` for a job board; property paths such as `/properties` or `/for-sale` and listing markup for real estate.
Schema.org types add 30 points, markup fingerprints 20, paths 12 and wording 8; a category needs 24 points, and a secondary category is kept when it reaches 35 percent of the primary one. A company site with a careers page can therefore be detected as jobs too, and receive these rules for that part of it.
The rules the listing module adds
A listing page is a crawled page whose URL contains `/listing`, `/listings`, `/annonce`, `/annonces`, `/property`, `/job` or `/jobs`, or one that carries `JobPosting`, `RealEstateListing`, `Offer` or `Product` markup.
On each listing page, the absence of any of those structured-data types is reported at medium severity. No seller, poster, agent or recruiter wording is reported at low severity. No form and no contact-seller, apply or contact wording is reported at medium severity, because a listing nobody can respond to is a dead end.
Across the site, the module looks for a search form, a search input, filter wording or sort controls, and reports their absence at medium severity. When listing pages exist, it also looks for pagination — `rel=next` or `rel=prev` links, `page=` parameters or `/page/` paths — and reports its absence at low severity.
When no listing page was reached, the module reports that listings could not be verified instead of inventing findings about pages it never saw.
What else changes, and for which of the four
Only the marketplace category is treated as a selling site. For it, the commerce checks weigh 1.3 times their base weight in the overall score, privacy, cookies and legal checks 1.2 times, performance, images and mobile 1.2 times, SEO and structured data 1.25 times; refund and shipping documents are expected by the legal discovery; `structure.cta` expects a purchase action; a FAQ is treated as relevant.
Marketplaces and directories both make an HTML sitemap relevant in `structure.html_sitemap`, since a large set of listings is exactly what a human-readable index helps with. For a marketplace, the monitoring check also points out that error tracking and real-user monitoring are usually worth having on a transactional platform.
Job boards and property portals get the listing rules and nothing else specific: their other checks run and weigh as on any site. The page does not claim more for them than the engine does.
Search, filters and crawling
Listing sites are where faceted navigation produces the most URLs: every combination of filters can become an address. The listing module only checks that search and filters exist; how many filter combinations a search engine should crawl is a separate decision, documented by search engines, that no automated rule makes for you.
The crawler follows internal links within its page budget, so on a site with thousands of listings the audit reads a sample. The listing findings describe the pages that were reached, and the report says how many that was.
Reading the listing findings
The listing rules run on every listing page the crawl reached, so a template problem — a missing seller line, no structured data — appears once per page built from that template. Grouped by rule, twenty identical findings are one fix.
Each finding records how it was detected: structured data for the markup rule, text analysis for the seller and contact rules, HTML parsing for search and pagination. Text rules look for English and French wording such as "seller", "vendeur", "posted by", "agent", "recruiter", "apply now" or "postuler", which is why a listing written in another language can be reported without a seller or a contact path it really has.
A "could not be verified" finding is emitted at information level. It says the crawl did not reach a listing page, and it never counts as a defect of the site.
Frequently asked questions
- Is a job board audited like an online shop?
- No. A job board gets the listing rules — structured data, poster information, a way to apply, search and pagination — but it is not treated as a selling site, so the commerce checks do not weigh in its score.
- Why does the audit ask for pagination?
- Because on a site with many listings, pagination is how visitors and crawlers reach the ones beyond the first page. The rule only runs when listing pages were found, and only looks for pagination links or page parameters.
- Which structured data counts on a listing page?
- Any of JobPosting, RealEstateListing, Offer or Product. The rule checks that one of them is present; whether the markup is complete enough for a particular search feature is a separate question.
Sources
Related
VeriFixScan crawls a site and applies its checks to every page it reaches, keeping the evidence behind each finding. Scanning one website is free.
Scan a website