Fashion Ecommerce Inventory: Managing Sizes, Colours & Seasonal Collections

A brand selling kitchen appliances might have 50 SKUs. A fashion brand with the same number of styles has 50 x 5 sizes x 4 colours = 1,000 SKUs. This variant multiplication is the fundamental challenge. Every inventory management decision, forecasting, reordering, allocating, marking down, has to be made at the variant level, not the style level.

The second unique challenge is time sensitivity. A kitchen appliance does not go out of season. A summer dress loses much of its sell-through potential once autumn arrives. This creates an urgency around markdown timing that does not exist in most other ecommerce categories. Mark down too early and you leave margin on the table (check the numbers with a profit margin calculator). Mark down too late and you are stuck with unsellable stock that occupies warehouse space for 9 months until next summer.

The third challenge is returns. Fashion has one of the highest return rates in ecommerce. Every return is a unit that was sold but is coming back, and the speed at which you process that return back to sellable stock directly impacts your available inventory for the next customer.

The most expensive mistake in fashion inventory is ordering equal quantities of each size. If you order 100 units each of XS, S, M, L, and XL, you will stockout on M and L while sitting on excess XS and XL, because size demand follows a bell curve, not a flat distribution.

A size curve is the percentage breakdown of how each size sells relative to total style sales. A typical UK women's fashion size curve might be: XS 8%, S 20%, M 32%, L 25%, XL 15%. That means if you are ordering 500 units total, the correct split is: XS 40, S 100, M 160, L 125, XL 75.

Critically, your size curve is specific to your brand and your customers. The industry average is a starting point, but after 2-3 seasons of sales data, you should be using your actual size distribution. The same applies outside clothing: a dog collar range, for example, may skew heavily towards Medium because so many popular breeds are medium-sized. Ordering equal quantities by size would leave it chronically short on Medium and overstocked on XS and XL.

Colours are harder to forecast than sizes because they are driven by trend cycles rather than demographic distributions. A size curve is relatively stable year-over-year, your customer base does not suddenly become 50% XL. Colour preferences can shift dramatically: a colour that sold 15% of your range last year might drop to 5% this year if the trend cycle moves on.

The practical approach: divide your colours into three tiers. Core colours (black, navy, grey, the colours that sell consistently every season regardless of trends) get 50-60% of your colour allocation. Trend colours (the current season's fashion colours) get 25-35%. Experimental colours (new colours you are testing) get 10-15%.

The same approach works for patterns: core patterns get the largest share of production orders, popular patterns from previous seasons get the next share, and new designs launching for the first time get the smallest. This tiered approach limits the downside risk on colours that do not sell while ensuring core bestsellers are always in stock.

Fashion ecommerce typically operates on a 4-season calendar: Spring/Summer (SS) drops February-March, sells through April-August; Autumn/Winter (AW) drops August-September, sells through October-February. Within each season, there may be sub-drops or capsule collections that create additional inventory management complexity.

The timing challenge is that production and ordering happen months before the selling season. SS production orders typically go in October-November for delivery in January-February. AW orders go in April-May for delivery in July-August. If you source from China with 190-day lead times, the calendar is even more compressed.

This means you are making purchasing decisions for your SS collection in autumn, when you have no current-season data. You are relying on last year's SS performance, adjusted for growth trends and new styles. The accuracy of these forecasts, especially for new styles with no sales history, determines whether you have a profitable season or one where you end up marking down 40% of your range.

Best practice: over-order on core/carry-over styles (you have historical data, the risk is low) and under-order on new styles (test with smaller quantities, then reorder via air freight if demand materialises). This limits downside risk while ensuring your proven sellers are in stock.

If your return rate is 30%, then for every 100 units you sell, 30 come back. These returned units are not lost sales, they are inventory that needs to be processed and returned to sellable stock as quickly as possible. The speed of returns processing directly impacts your weeks cover accuracy and available inventory.

If you process returns within 24-48 hours, those units are back on your shelves and available for the next customer. If returns sit in a pile for a week, your system shows lower stock than you actually have, which might trigger unnecessary reorders or prevent you from selling units that are physically available but not system-available.

Fashion return reasons vary, and each has different inventory implications. "Wrong size" returns usually go straight back to sellable stock. "Not as described" or "quality issue" returns need inspection and may need to be written off. "Changed mind" returns go back to sellable stock but may need repackaging.

Accessories such as dog collars usually see fewer returns than clothing because there is less fit uncertainty, but returns still need processing. A dedicated returns location in Shopify keeps returned stock separate from sellable stock until inspection is complete.

Markdown discipline separates profitable fashion brands from those that haemorrhage margin every season. The rule: if a style is not meeting its sell-through rate by the midpoint of its selling season, mark it down. Do not wait until end of season when everyone else is discounting and competition for discount-seeking customers is fiercest.

A practical markdown cadence for fashion ecommerce:

Weeks 1-4 of season: Full price. No discounts. Measure initial sell-through.

Week 5-6 (midseason check): Review sell-through rates. Any style below 40% of expected sell-through gets flagged.

Weeks 7-8: First markdown: 20-25% off on flagged styles. This catches early discount shoppers.

Weeks 9-10: Second markdown: 40-50% off on remaining flagged stock. Priority is clearing units, not maximising margin.

Week 11+ (end of season): Final clearance: 60-70% off. The goal is to recover whatever cash you can. Carrying unsold seasonal stock into the next year costs more in warehousing and opportunity cost than selling at a steep discount.

The common mistake: holding stock hoping it will sell, then being forced into a single massive markdown at end of season. Progressive markdowns recover more total cash because different customer segments shop at different discount tiers.

Fashion inventory demands specific capabilities from your inventory management tool:

Variant-level tracking: not just product-level. You need weeks cover per size per colour, not averaged across the style.

Size curve analytics: the tool should show you your actual size distribution so you can order proportionally.

Sell-through rate monitoring: real-time sell-through as a percentage of initial stock, per variant, with alerts when rates fall below threshold.

Seasonal lifecycle management: the ability to mark styles as seasonal and track their lifecycle from launch to clearance.

Returns integration: returned units updating available stock in real-time, with quality categorisation.

Canopy Inventory, a read-only Shopify app coming soon to the Shopify App Store, covers part of this list. It works out weeks of cover for every SKU, so each size and colour is measured separately rather than averaged across the style, and it gives a suggested buy quantity per SKU based on that SKU's own sales. It does not provide size curve analytics, seasonal lifecycle management or returns processing.