The Hidden Cost of Too Many SKUs

Every additional color, size, material, pack, or channel-specific version can create a distinct item to identify and operate. The visible cost is the inventory purchased. The less visible cost is the work required to receive, count, photograph, describe, locate, replenish, reconcile, discount, return, and eventually retire that item.

This does not mean fewer SKUs are always better. A size range, color choice, replenishment program, or channel pack can be essential to the customer and the business. The useful question is whether each variation has an accountable role and whether the operating system can support it accurately.

Define the item before counting it

“Style count” and “SKU count” are not interchangeable. A style offered in five sizes and four colors may require twenty sellable combinations before packs, markets, or other distinctions are considered.

GS1’s specifications apply unique identification rules to apparel and home fashions. The standard describes a one-to-one relationship between an individual trade item, color, size, and its identifier in apparel examples. The GTIN Management Standard provides rules for when a trade-item change requires a new identifier. These are supply-chain identification rules, not a mandate for how a small label must name internal records, but they demonstrate why casual variant proliferation becomes operationally consequential.

Create a data dictionary for style, color, size, material, season, pack, channel, location, and status. State which combination forms the internal SKU and which items receive external identifiers. Preserve predecessor and successor links when an item changes.

Inventory value is only the first line

At a minimum, calculate on-hand units multiplied by a documented unit-cost basis. IRS Publication 334 explains that businesses producing, purchasing, or selling merchandise generally need an inventory method that clearly reflects income, with special rules and exceptions for qualifying small businesses. It also distinguishes raw materials, work in process, finished goods, and supplies that become part of the product.

That general tax guidance does not tell a team what carrying-rate percentage to use. Carrying rates often combine capital, storage, insurance, handling, shrink, obsolescence, and opportunity assumptions. List the included components and their owners rather than importing an unexplained industry percentage.

Keep financial inventory valuation separate from a management estimate of complexity. The first needs consistent accounting treatment. The second can surface labor and risk that the accounting inventory value does not attempt to allocate per SKU.

Count the recurring touches

A SKU can trigger work at several points:

• setup and identifier assignment;
• product specification and bill-of-material maintenance;
• sample, label, image, description, and accessibility review;
• purchase-order and supplier communication;
• receiving, put-away, location assignment, and discrepancy resolution;
• cycle counts, transfers, replenishment, and stock adjustments;
• return inspection and disposition;
• markdown, archival, and end-of-life decisions.

Estimate minutes only when the task and measurement method are defined. An hourly rate in a planning model should be a lawful loaded-cost assumption owned by finance or operations, not a worker-speed target. Time spent correcting bad data belongs in the ledger too.

Measure demand fragmentation

Adding a variation can spread the same demand across more records. A top-level style may appear healthy while several size-color combinations remain stranded. Review sell-through, days or weeks of supply, stockout frequency, aged units, return reasons, and replenishment separately by the decision unit.

Shopify’s current inventory reports document product sell-through, inventory snapshots, inventory remaining, and ABC analysis. Those fields can help a merchant form questions, but platform reports must be reconciled to the business’s tracking settings, negative inventory behavior, locations, returns, transfers, and reporting period. An ABC class is not an automatic delete instruction.

Census inventory-to-sales data offers a broader retail context with published reliability information. It is not granular enough to validate a brand’s assortment and should not be used as a target for one company.

A reproducible fictional complexity ledger

FashionMember created six invented SKU rows in content/data/FM-132-sku-complexity.csv. The script scripts/fm132-sku-complexity.php calculates five illustrative components:

inventory value × user-supplied carrying rate

surplus units × unit cost × user-supplied markdown-loss rate

plus receipt labor, count labor, and content-maintenance labor based on invented minutes and hourly rates.

SKU-A holds $6,160 of fictional inventory and produces an illustrative annual complexity cost of $1,656. SKU-B has 480 units on hand against 200 invented annual demand units; its 280-unit surplus and other assumptions produce $4,894. SKU-D combines a smaller inventory value with a larger fictional surplus and more counting and content touches, producing $2,741.60.

The six-row fixture totals $12,113.57, but that number is not a cost benchmark. SKU-E is held because its fictional data is marked open and lacks a human owner, even though the script can calculate $1,140.43. Arithmetic does not repair provenance.

The model excludes shared creative work, lost sales from stockouts, size availability, customer choice value, supplier minimums, pack constraints, channel strategy, return behavior, taxes, financing, and correlations among items. Its carrying and markdown rates are invented inputs, not FashionMember recommendations.

Use a keep, repair, test, or retire review

Do not rank items from one blended score. Give the merchant four possible review states:

• keep: the item has a clear customer and assortment role, reliable data, and supported operations;
• repair: identity, imagery, description, location, or replenishment data is causing avoidable friction;
• test: the role is plausible but the demand or service evidence is incomplete, so define a limited period and success criteria;
• retire: an accountable owner approves the end-of-life plan after checking customers, commitments, stock, returns, warranties, identifiers, and channel obligations.

Document qualitative reasons. A low-volume extended size may be strategically necessary and should not be judged by the same rule as a redundant color. A spare or repair item may have service value. A new item may not yet have enough history for a stable rate.

Add governance before adding variants

Require a short variant brief before creation: customer need, incremental demand hypothesis, cannibalization risk, minimum, identifier, content burden, storage location, replenishment rule, review date, and retirement owner. The burden should be proportional; the objective is clarity, not bureaucracy.

Then audit the item master. Look for duplicate records, missing attributes, reused identifiers, inconsistent status, orphaned images, negative inventory, unresolved transfers, and products that remain purchasable after retirement. Correct the source system rather than building a permanent reconciliation spreadsheet around bad data.

The hidden cost of a SKU is therefore not proof that choice is wasteful. It is evidence that assortment architecture and operating architecture must be designed together. A variation earns its place when the customer purpose is explicit and the business can identify, support, measure, and retire it responsibly.

Sources and verification

• GS1 standards repository — official repository for the GTIN Management Standard, General Specifications, product-image standards, and related identification standards.
• GS1 General Specifications — official current standard, including apparel and home-fashion identification considerations.
• GS1 GTIN Management Standard — official rules for consistent decisions about unique trade-item identification.
• IRS Publication 334 — official general small-business discussion of inventories, methods, cost of goods sold, and gross profit; not company-specific tax advice.
• Census Monthly Retail Trade Inventories — official retail inventory and inventory-to-sales files with adjustment and reliability information.
• Shopify inventory reports — current first-party definitions and limitations for inventory report fields.
• Shopify ABC inventory analysis — current first-party explanation of its product classification workflow; not an independent assortment standard.