Turning a catalog-wide SEO problem into an auditable system
A five-figure furniture catalog needed more than page edits. The work connected catalog integrity, search metadata, product templates, and repeatable implementation.
01 / situation
The catalog had outgrown page-by-page SEO.
A large furniture store carried products from many vendors, each arriving with different structures, labels, descriptions, and inventory behavior.
Organic discovery, product templates, shopping feeds, and operations were all consuming the same inconsistent catalog data. Optimizing isolated pages could improve individual records, but it could not stop the next import or update from recreating the problem.
02 / evidence
The audit made the problem countable.
The first useful output was not a recommendation deck. It was a normalized view of the catalog and an issue taxonomy that could be queried and acted on.
| Evidence class | What it exposed | Why it mattered |
|---|---|---|
| Search metadata | Missing and inconsistent titles, types, and tags | Weak relevance and discovery signals |
| Catalog integrity | Duplicate identifiers and orphaned records | Unsafe updates and feed collisions |
| Product content | Missing descriptions and media gaps | Thin product experiences |
| Storefront dependencies | Theme fields without dependable sources | UI promises the data could not fulfill |
SOURCE NOTE / Dated catalog audit and theme dependency map. Client identity and commercially sensitive values are withheld.
03 / diagnosis
SEO quality was downstream of catalog integrity.
The constraint was an unreliable data layer feeding metadata, navigation, templates, Merchant Center, and repeated vendor operations.
The work needed one source of truth, issue classes with different recovery paths, and an implementation process that could prove what it changed.
04 / intervention
Turn the catalog into a system that can be inspected.
The intervention connected audit, prioritization, field mapping, and safe mutation rather than treating each as a separate project.
- Normalize source data before judging records.
- Separate critical integrity failures from enrichment opportunities.
- Map storefront fields to the exact theme elements that consume them.
- Verify platform definitions and types before any bulk write.
- Use smoke tests, skip-existing behavior, and per-record logs.
05 / systems
Automation supported the SEO plan.
Vendor-specific parsers and a universal fallback converted structured information already present in product content into fields the storefront could use.
The tooling was deliberately reversible. It staged or skipped uncertain records, preserved populated values by default, and left a JSONL trail that could be audited after a run. The software is evidence of implementation depth—not a different offer competing with Shopify SEO.
06 / verification
Every layer had a check.
Verification covered the audit dataset, theme-source map, parser coverage, platform field definitions, dry runs, smoke tests, and per-row mutation logs.
- Source dates and catalog scope attached to audit outputs.
- Before/after values retained for changed records.
- Exceptions withheld instead of silently coerced.
- Public case language distinguishes tested coverage from completed live population.
07 / result
An unbounded problem became an operating plan.
The work transformed a vague page-by-page SEO backlog into exact issue classes, mapped dependencies, and a safe path for catalog-wide improvement.
Verified result: an auditable catalog system and reversible implementation path. Performance outcomes remain private until a matched baseline, window, source, and publication approval are available.
08 / lesson
Large-store SEO begins below the page.
For complex Shopify stores, the fastest durable improvement often starts with the data architecture that produces pages—not another isolated round of edits.
That lesson now shapes how I approach audits: inspect the catalog and operating workflow early, because they determine whether page-level recommendations can survive contact with the real store.