This audit is generated directly from the client's product catalog by the fitment engine — every number below is computed, not estimated. No changes were made to the store to produce it.
| Measure | Value |
|---|---|
| Products in catalog | 2,310 |
| Categories in catalog | 1,779 |
| Fitment pages present | 1,682 |
| Products the engine can place with confidence | 2,283 of 2,293 (99.6%) |
| Missing fitment pages | 0 |
| Fitment pages missing SEO title or description | 0 of 1,682 |
| Products absent from their fitment pages | 0 |
| Empty fitment pages (doorway-page risk) | 42 |
| Machines competitors cover with no page on this store | 139 |
| Naming conflicts needing a human decision | 4 |
| Products with name defects (still parseable) | 41 |
| Products flagged for review | 10 |
Machines the products fit, with no page for a buyer (or an AI shopping agent) to land on.
Fitment pages with no page title and/or meta description — pages that exist but tell search engines nothing about what they are.
Products whose name identifies the machine they fit, but which are not assigned to that machine's page — invisible exactly where buyers look.
Category pages with zero products anywhere beneath them. Google's spam policy treats pages "created to funnel users" with no standalone value as doorway pages; empty fitment pages are the textbook case and can drag down the whole subtree. The first ten of forty-two:
Machines where a competitor publishes a page whose URL itself names the make and model, and this store has no matching fitment page. The method is deliberately conservative — a competitor only earns a machine from its own URL, never from marketing copy or search snippets. The first eight of one hundred thirty-nine:
Naming conflicts (4) — same page, two spellings; needs a human call:
ASV › RT65 Max vs existing ASV › RT65 MAXASV › RT65 Max › Tracks vs existing ASV › RT65 MAX › TracksTakeuchi › Tl120 vs existing Takeuchi › TL120Takeuchi › Tl120 › Tracks vs existing Takeuchi › TL120 › TracksName defects parsed anyway (41) — missing spaces or stray hyphens the engine corrected; worth fixing at the source:
numeric_variant_suffixnumeric_variant_suffixglued_part_keywordglued_part_keywordmissing_space_before_dashFlagged for review (10) — the engine refuses to guess on these:
no_make_matchno_make_matchno_make_match; no_part_keywordEvery gap above is a machine-specific search — "Bobcat E32C tracks", "ASV PT75 rubber tracks" — where the store either doesn't rank or shows buyers a broken experience, and AI shopping agents reading the catalog structure hit the same walls. The engine fixes all four gap classes with native catalog objects: real categories, real assignments, real metadata, no injected pages. Each change is planned, applied, and verified idempotently, journaled with rollback.
Deterministic extraction against the store's own naming conventions; page creation gated by confidence scoring — pages are only proposed where product evidence supports them; conflicts and unparseable products are surfaced for human decision, never guessed. The audit is read-only.
Fixed fee by catalog size — $2,400 up to 10,000 products, $4,800 to 50,000, quoted above that. Read-only access, report in two business days.
Request an audit