Resolution Layer Case Study · Read & Diagnose
A national sporting-goods retailer.
A multi-lens data-quality read of the catalog — and where it leads.
This is a real EKOM catalog analysis, with the retailer's identity removed. The client is a national sporting-goods retailer — footwear, apparel, and equipment across the major athletic brands. EKOM ran an automated, multi-lens quality analysis over the retailer's live product catalog and surfaced 622 distinct data-quality issues — the kind standard validation misses, because it checks whether a field is filled, not whether the value is right.
What follows leads with the findings that carry real risk — the wrong team on team merchandise, a price that undercuts its own "original," a jersey filed where no shopper will find it — then groups the rest by theme, with real examples throughout. Third-party brand names are kept as they appeared; only the retailer's own identity has been removed.
These are machine-surfaced findings meant as a triage signal, not a verified defect list — a small share may be intentional. Even so, the concentration and specificity of what surfaced is a strong read on catalog health.
What's inside
- The findings that carry real risk — the highest-impact issues a shopper sees, or that quietly keep them from finding the right product.
- Operational & data-integrity issues — the less-visible defects that create risk in fulfillment, routing, and compliance.
- The SKU layer — a distinct class of variant-level issue the product view can't see.
- How EKOM reads a catalog — why category-specific problems surface here instead of averaging away.
- What it means, and what's next — from diagnosis to an enriched, channel-ready catalog.
Where the issues concentrate — by product area, and by theme. Many issues span more than one theme, so theme counts are approximate.
The count isn't the point. These are issues that pass every completeness check — the field is filled, so a validator waves it through — yet are wrong in ways that only surface when structure and meaning are read together. They cluster where they hurt most: discovery, pricing, and fulfillment — and they compound as the catalog scales.
By product area
| Product area | Issues |
|---|
| Apparel | 353 |
| Footwear | 195 |
| Accessories | 25 |
| Licensed / fan gear | 14 |
| Equipment | 12 |
| Cleats | 11 |
| Mixed / other | 12 |
By theme
| Theme | ~Count | Who feels it |
|---|
| Naming, spelling & product-type errors | ~166 | Shoppers |
| Missing identifiers / core catalog fields | 111 | Shoppers + ops |
| Stale or contradictory status / availability flags | 93 | Shoppers + ops |
| Wrong categorization / navigation | 89 | Shoppers |
| Wrong filterable attribute values | 57 | Shoppers |
| Internal codes / HTML / spreadsheet artifacts in live copy | 49 | Shoppers |
| Wrong or missing licensed-team tags | 35 | Shoppers |
| Contaminated / wrong copy | 20 | Shoppers |
| Duplicate values | 2 | Shoppers |
The findings that carry real risk.
Highest impact — the issues a shopper actually sees, or that quietly keep them from finding the right product.
Highest-Visibility Finding · Licensed / Fan Gear
The wrong team on team merchandise.
On a team-loyal storefront, these are the ones that get noticed: an Atlanta Braves tee filed under the Los Angeles Dodgers team shop; a Milwaukee Bucks snapback whose detail bullets read "Team: Chicago Bulls"; an Indiana Fever (WNBA) jersey tagged as the Indiana Pacers (NBA); a Las Vegas Raiders hat still listing "Oakland Raiders". And every Pro Standard MLB licensed item has a blank team field — so none of them appear in shop-by-team navigation at all.
Products in the wrong place on the site
Shoppers can't find items, or find the wrong ones: a jumpsuit misclassified as a pant (routed to pants, absent from where it belongs); a men's soccer shorts item classified under women's — a pattern that recurs across several men's/women's items; several live, searchable items with no category at all, orphaned from navigation entirely; three jacket-vest items routed to the "T-Shirts" category; and a skort typed as a "short."
Wrong or unprofessional copy on the product page
A Louisville Slugger softball bat whose description carries Air Jordan editorial copy — with an unrelated hyperlink — bled across from a completely different product. The internal vendor hold status "TRUE RELIGION (HOLD)" shown to shoppers. Internal merchandising codes pasted verbatim into customer-facing descriptions. A live The North Face jacket with an unfilled template placeholder "- COLOR" in its product name. And bookmark <span> tags from a rich-text editor embedded in live description fields, corrupting the page HTML.
Naming & spelling errors
Live on the site: "Demin" (Denim), "Sim Fit" (Slim Fit) on track jackets, "Woman's" for "Women's", "Shortse" (product type merged with a stray word), and "TNew York Mets". A Nike Women's running vest named "Nike Men's…" and a women's shoe labeled "Men's" — despite every other field confirming the correct gender. And "Under Under Armour", the brand name duplicated in a legging's description.
Wrong attributes that break filters
denimWash populated on non-denim items; a 100% polyester jersey tagged "Denim." Non-fleece joggers tagged "Fleece"; cycling shorts tagged "Tights" (so they vanish from shorts filters and appear under tights); a crew-neck top tagged "Half-Zip" with no zipper. And missing sizeChartId on many items, so the size-chart widget never renders — a known driver of returns, especially in kids' apparel.
Operational and data-integrity issues.
Less visible to shoppers, but they create real risk in fulfillment, routing, and compliance.
Missing core fields
webid (the web identifier — a blank can break product-page URLs); manufacturer blank on items with a known vendor; inconsistent vendorPartNumber formatting; and missing size-chart IDs. These are the fields other systems quietly depend on.
Guardrail Contradiction
Items that failed the catalog's own quality gates — still live and buyable.
Several items are failing quality guardrails yet remain searchable and purchasable. One carries a bypassGuardrails flag set True with no documented reason. Others are marked offline for "no image" while online with several images, and one is offline but still flagged in-stock and purchasable. The status layer and the availability layer disagree with each other on the same products.
Duplicate / repeated content
Duplicated detail bullets that make a page look broken, and duplicated fabric-composition lines — small, but they read as carelessness on an otherwise premium product page.
Beyond the storefront: the SKU layer.
Everything above is at the product level. EKOM also ran a variant (SKU) analysis — one row per individual size and color a shopper can actually buy — on a sample of the catalog. It surfaces a different class of issue the product view simply can't see, because the product view summarizes its variants down to a count.
- Price integrity. Items where the current price is higher than the "original" — a fake sale, and factually wrong — and active items priced at $0.00 against a normal original, a free-item checkout risk.
- Barcodes / UPCs. Blank or 0 UPCs on live items (nothing scannable for point-of-sale or EDI), and UPCs that fail GS1 UPC-A validation in vendor routing.
- Shipping-breaking dimensions. The primary variant of nearly every product has height and width set to 0.0 — which breaks shipping-rate calculation and carrier-manifest validation.
- Per-variant availability. Every size of a purchasable Jordan Luka 5 "Venom" is flagged non-searchable (invisible to on-site search); other live items carry a blank inventory count.
- Broken product grouping. Six Sabrina 3 "Silencer" size variants point to the wrong parent product (grouped under an unrelated item); some legacy variants have no parent at all.
- Within-product inconsistency. Sibling variants of one product with mismatched names or tax classes — and nine Nike men's SKUs labeled "Women's Shoe."
Scope. This was a sampled look at four product groups — a proof of the variant layer's value, not an exhaustive pass. A full variant analysis across the catalog is a straightforward next step.
How EKOM reads this catalog.
Why category-specific problems surface here rather than averaging away in a whole-catalog scan.
1
Profile the catalog
EKOM profiled the catalog with no schema supplied — inferring the vertical and every field's role, fill rate, and quality signals before any analysis ran.
No manual setup required.
→
2
Partition & pass
It then ran 48 focused passes — partitioning the catalog by product family (apparel, footwear, licensed, equipment, cleats…) and, within each, by sub-type or brand. Each pass examines one coherent slice.
One coherent slice at a time.
→
3
De-duplicate & group
Findings are de-duplicated across passes, then grouped by theme and impact — which is why issues like team tags on fan gear, denim attributes on jeans, or bra-support on bras come through cleanly.
Signal, not noise.
A single whole-catalog sweep averages a large book toward its mean and loses the per-category defect. Focused passes are how a men's shorts item filed under women's, or a licensed team field left blank, gets caught instead of buried — and it's what makes the next step precise: EKOM knows exactly which fields need filling and which are already sound.
What this means — and what's next.
more than25%
More than one in four organizations facing data-quality challenges report losing over $5 million annually as a result — and that figure doesn't account for the compounding impact on digital commerce, where every search filter, variant selector, size-chart lookup, and category page relies on clean structured data to function. In retail, catalog structural failures don't announce themselves. They disappear quietly into broken filters, empty collection pages, and shoppers who couldn't find what they were looking for.
Forrester Data Culture and Literacy Survey, 2023 (RES181258)
This retailer's digital catalog is in an active phase of expansion — a dedicated kids storefront, continued investment in the platform that manages product content, and an order-management rollout across a large national store fleet. Each of those investments performs at the level the underlying data allows. The 622 findings in this analysis represent the gap between where the catalog data is today and where it needs to be to support what's already been built on top of it.
None of these are content problems. A jersey filed under the wrong team is invisible to shop-by-team navigation no matter how well it's described. A missing sizeChartId silently prevents a size chart from loading — the kind of friction that ends a session. A price that sits above its own "original," or a variant with height and width of zero, is a revenue-and-fulfillment risk no content system can see. These accumulate invisibly because standard validation checks for presence, not correctness. EKOM's catalog resolution layer checks both — reading each category in full.
This pass read and diagnosed. The same structural understanding powers the work that follows — turning a diagnosed catalog into one that holds up through every peak season and expansion ahead.
1 · Apply the confirmed corrections
The wrong-team tags, miscategorizations, naming/spelling errors, filter-breaking attributes, and the guardrail/status contradictions — resolved systematically from data already in the catalog, not one record at a time.
2 · Enrich for discovery
Fill the fields that make products findable and complete — sizeChartId so size charts render (a known returns driver), manufacturer and team tags for navigation, and the filterable attributes tuned for on-site search — so products show up where shoppers look for them.
3 · Go to the SKU layer, and hold the line
Run the full variant-level pass (price integrity, UPCs, per-variant availability, shipping dimensions) as the natural next depth — and keep ongoing intelligence at intake so a catalog scaling under new investment stays clean instead of re-accumulating these gaps.
This is how EKOM moves a catalog from insight to impact —
and keeps it right as the business grows.
About this analysis
This is EKOM's second pass on this retailer's catalog. The first, in an earlier engagement, examined a sample across the retailer's two storefronts. This pass read the catalog with no schema and no manual setup — the pipeline profiled the catalog on its own, identifying the vertical and every field's role before any analysis ran. It then read the catalog through 48 focused passes, each examining one coherent slice, so category-specific defects surface cleanly instead of averaging away. The same resolution, now across the entire book, with nothing built by hand.