Duplicate SKUs and Pricing Problems: How to Detect, Merge, and Prevent Them

A buyer created a SKU for an MRO bearing under the manufacturer part number. Three months later, a supplier feed onboarded the same bearing with a slightly different description and no MPN — and your system created a second record. Now the catalog carries both, each with its own cost, its own list price, and its own sales history. This is the core of duplicate SKUs pricing problems: when one real product lives behind two or more records, every downstream number that depends on it drifts silently out of alignment.

Claro resolves this at the identity layer. As supplier feeds, PIM imports, and manual adds arrive, Claro matches each incoming record against your existing catalog using fuzzy scoring across part numbers, descriptions, and identifiers — catching near-duplicates before they land. When duplicates already exist, Claro merges them into a single golden record, reconciles cost and price, and writes the clean record back into your PIM or ERP. The catalog stays trusted as it grows, rather than degrading with each new feed.

How duplicate SKUs break pricing

When two records represent one product, pricing logic has no single source of truth. A repricing rule keyed on cost margin updates the record it can see and leaves the duplicate stale. Customers then encounter two prices for the same item — on the same site, in the same quote, sometimes in the same cart.

The cost side is worse because it is invisible. A CPG distributor buying a cleaning concentrate under two SKUs may have created one record when landed cost was lower and another after a later freight surcharge. Average cost, last cost, and standard cost now diverge per record, so margin reports compute against the wrong baseline. You believe you are clearing 22 % on the item when half the volume is actually moving at 14 %.

How duplicates corrupt analytics and demand signals

Pricing is downstream of demand, and demand is exactly what duplicates fragment. A furniture retailer selling the same dining chair under two SKUs splits its unit velocity in half. Each record looks like a marginal performer, so both get deprioritized in merchandising, demoted in search, or flagged for discontinuation — while the true combined demand would have earned premium placement.

The same fragmentation distorts inventory and forecasting:

For an industrial distributor, understated velocity means safety stock set at half the real demand — leading to stockouts on exactly the items customers order most. The forecast was never wrong about the market; it was wrong about the catalog.

Why duplicate SKUs are so hard to spot

Duplicates survive because they are not identical. They are near-matches: the same product with a transposed manufacturer part number, a different unit of measure (each vs. case), a vendor-supplied description versus an internal one, or a UPC on one record and none on the other. Exact-match queries miss all of these, which is why teams consistently underestimate how many they have until they measure properly.

• Same product under different MPN formats (dashes, spaces, leading zeros)
• Each vs. case vs. pack records that should roll up to one parent SKU
• Supplier-feed records duplicating items already in the catalog
• Legacy SKUs left active after a PIM or ERP migration
• Variants (color, size) modeled as standalone products with their own pricing

Fuzzy matching across multiple normalized attributes — part numbers, descriptions, and identifiers scored together — is the only reliable way to surface these. That detection step is the bridge to a durable fix: Claro’s canonical product layer resolves every incoming record to one identity at intake, so a duplicate cannot be created in the first place.

Before and after: messy catalog vs. trusted catalog

A practical path to clean pricing

1. 1
   Measure the overlap

   Quantify how many records collapse into how many real products before touching pricing. The ratio reveals how distorted your current numbers already are — and sets the baseline for measuring improvement.

2. 2
   Score candidates, not exact matches

   Use fuzzy scoring across MPN, GTIN, description, and category so near-duplicates surface instead of hiding behind formatting differences. A duplicate SKU finder or confidence-scored matching pass handles the cases that exact-key queries miss.

3. 3
   Merge into a golden record

   Consolidate each cluster into one canonical product record with reconciled cost, price, and identifiers. Keep merges reversible and auditable so you can unwind any incorrect consolidation without data loss.

4. 4
   Write the clean record back

   Push the golden record back into your PIM or ERP so every downstream system — pricing engine, analytics, e-commerce feed — reads from one trusted source. Claro handles this write-back step natively, preserving your existing system architecture.

5. 5
   Prevent re-entry at intake

   Resolve every new supplier feed and manual add against existing records before they enter the catalog. If an incoming item matches an existing product above the confidence threshold, it attaches to that record rather than creating a new one.

Related

FAQ