Peak season returns: the size problem to fix before Q4

Why peak season returns spike — and why it isn't the gift wrapper

Peak season returns are not a January logistics problem. They are a size problem, and the cheapest time to fix one is in September — before the first pallet leaves, not after the boxes come back.

Retailers expected roughly 17% of holiday sales to be returned in the 2025 season, according to the NRF and Happy Returns 2025 Retail Returns Landscape report, and online orders ran higher still at 19.3%. The standard advice — hire seasonal staff, extend the window, renegotiate reverse-logistics rates — all manages a return you have already paid for. It never stops the return from being created.

The one cause you can still remove before the peak is the buyer who could not tell the size before ordering. For furniture, industrial parts and building materials, that is most of the spike.

What peak season returns actually are

Peak season returns are the share of holiday-period orders that buyers send back, and they run above the annual average because gift buyers never inspect the product in person and the return window is longer. In 2025 the holiday rate sat at about 17% of holiday sales, against 15.8% of total annual sales and 19.3% of online sales.

Two mechanics separate the peak from the rest of the year:

The chooser is not the keeper. A gift is picked by one person and lived with by another, so a "close enough" size becomes a return.
The window is longer. On Amazon, most items bought between 1 November and 31 December 2025 could be returned until 31 January 2026 — an extra month for the buyer's decision to change (Amazon's holiday return policy).

A long window plus gift uncertainty turns a size question a showroom would have settled in thirty seconds into a return that lands in your data eight weeks later. And the online channel keeps growing: U.S. e-commerce reached $340.2 billion in Q2 2026, up 12.2% year over year, and now accounts for 17.1% of total retail sales (U.S. Census Bureau, Quarterly Retail E-Commerce Sales). More first-time online buyers means more size guesses.

Why the size question is the expensive one

Most returns cost you on the way back. A size return costs you twice: the buyer was never a candidate to keep the item, and a used oversized item is often hard to resell as new.

Baymard Institute found that 42% of shoppers judge size from the product images, and that 28% of e-commerce sites offer no in-scale image at all (Baymard Institute — All Products Need at Least One 'In Scale' Image). With no scale reference, the buyer does the only thing left: orders two sizes and returns one. NRF found close to two-thirds of consumers admit to at least one costly returns behaviour — bracketing among them.

Return driver	What the buyer is looking at	Who absorbs the cost	Fixable before peak?
Wrong size / doesn't fit	A photo with no dimensions	You: return freight, restocking, markdown	Yes — a measured, in-scale image
"Not as described"	A spec that disagrees with the goods	You: dispute, chargeback, rework	Yes — one measured spec source
Damaged in transit	A carton that failed handling	You and the carrier	Partly — carton dimensions and spec
Bracketing (buy two, keep one)	Unclear size, so the buyer hedges	You: double handling	Partly — a clearer size reduces it
Changed mind	Nothing wrong	You: reverse logistics	No — policy only

Only two rows are fully in your control before the season starts, and both are about measurement, not logistics.

Scenario A: You supply furniture to overseas retailers

Your buyer is a retailer, but the return is triggered by their customer. A wardrobe that "won't fit the alcove" or a sofa that "looks smaller than the photo" comes back as a claim against your invoice — even when your unit dimensions were correct. The gap is almost always between what you measured and what the listing showed.

Fix first: publish one measured spec diagram per SKU carrying the dimensions a buyer actually needs to judge fit — overall, internal, and the clearance the item requires. Agree the baseline on a standard reference, then use a furniture spec diagram format your retailer can drop straight into their product page. Put the packed carton dimensions on the same sheet so the retailer's warehouse and the carrier agree with you too.

Scenario B: You sell on marketplaces

On Amazon, TikTok Shop, Walmart and Etsy the holiday window stretches past the new year, so a December order can come back in late January. Every image in the stack has to answer "how big is this?" without a salesperson. The category data is blunt: size-driven returns cluster at the top in furniture and apparel and sit far lower in electronics — the full spread is in size return rate by category.

Fix first: at least one in-scale image with the measured dimensions printed where the eye lands, plus a size table in the description that repeats the same numbers. If the image and the text disagree by even a centimetre, the buyer believes the image — and returns the order.

Scenario C: You export building materials

Cut-to-size panels, tiles and profiles fail on the gap between nominal and actual size, and on coverage per box. A buyer who expected 1.44 m² of tile from a box and gets 1.35 m² does not file a "size question" — they file a short-shipment claim.

Fix first: label coverage per box and the manufacturing tolerance on the packing spec, not just the catalogue. A spec diagram that shows nominal size, actual size and tolerance in one view ends the argument before it starts.

Which fix to ship first

Your situation	The return you actually get	Fix first	Asset to ship
Furniture supplier → overseas retailer	"Doesn't fit the space" claim	Measured overall + internal dims	One spec diagram per SKU
Marketplace seller	Size return, bracketing	A single in-scale dimension image	Dimension-labelled main image
Building-materials exporter	Coverage / nominal-vs-actual dispute	Coverage per box + tolerance	Dimensioned packing spec
Industrial parts supplier	Wrong-fit or wrong-thread return	Thread, pitch and fit callouts	Measured part diagram

The pre-Q4 checklist

Pick the 20% of SKUs that generate the most size questions — not all of them
Measure each one once and store the numbers as the single source
Put overall, internal and clearance dimensions on a diagram, not in a paragraph
Export one in-scale image at the pixel size each marketplace expects
Add packed carton dimensions to the same spec sheet for freight and warehouse
Make the description table repeat the diagram numbers exactly
Re-check every SKU against the platform's current image rules before the peak

What "measured" has to mean

The fix is not "take better photos." It is locking the actual measured dimensions onto the image where buyers look — a measurement snapped to the real edge of the product, then exported at the spec-diagram size each platform wants. That is a different job from a photo editor that lets you draw an arrow and type any number beside it, and a different job again from an AI image generator that prints a number that looks plausible but was never measured. A label is only as trustworthy as the measurement behind it. When the number comes from geometry rather than from a guess, the listing, the spec sheet and the carton all inherit the same figure — which is exactly what stops the "not as described" claim.

Next steps

Pick the path that matches your situation:

If returns are already painful, run the numbers before you spend anything — the return cost calculator turns a return rate into a currency figure you can compare against the cost of re-shooting a SKU set.
If you have never costed the problem, start with the hidden cost of ecommerce returns so the fix gets budget, not a shrug.
If you need the asset itself, a diagram built on measured geometry — snap-to-edge measurement, one export per platform size — gets you there without a design retainer.
If you only have time for one thing, do the in-scale image. It is the single highest-leverage asset for size-driven returns.

FAQ

How much of peak season actually gets returned?

About 17% of holiday sales were expected to come back in the 2025 season, against 15.8% of total annual retail sales and 19.3% of online sales, according to the NRF and Happy Returns 2025 Retail Returns Landscape report. The holiday rate runs higher than the annual rate because gift buyers cannot inspect size in person.

Why do returns go up during the holidays?

Two reasons: the person choosing the item is usually not the person keeping it, and the return window is longer — Amazon allowed most items bought from 1 November to 31 December 2025 to be returned until 31 January 2026. Both stretch the time between a size guess and the return.

How do I reduce peak season returns without hurting sales?

Remove the size guess instead of the sale. A measured, in-scale image lets a buyer confirm fit before ordering, so the order that never had a chance of being kept does not get placed. Baymard found 42% of shoppers judge size from images and 28% of sites give them no scale reference — that gap is the opportunity.

Can an AI-generated product image show accurate dimensions?

Only if the number is measured, not invented. An image generator can produce a convincing label with a plausible figure, but the value is generated rather than measured. For anything a buyer will order against, the dimension has to come from the real geometry of the product.

When should I start preparing listings for peak season?

Now. The assets take time to measure and export across a catalogue, and a post that goes live in November competes with every other seller doing the same thing. Preparing in September and October means the size answers are already on the page when the traffic arrives.