17% of holiday sales come back: answer returns inside the clip
NRF's 2025 report expects 17% of holiday sales to be returned. Here is a four-clip answer set per product, cut from one master for about $0.38 on Sume.

The National Retail Federation's 2025 returns report says retailers expect 17% of holiday sales to come back, against 15.8% of annual sales and 19.3% of online sales. If a return starts with a question the product page never answered, the cheapest fix is a short clip that answers it. On Sume you can cut four such answers from one master recording for $0.08 in trims, then join and caption them for $0.30 more.
The figures worth planning around
These are the return numbers from NRF's 2025 release. They describe the whole retail market, not your store, so use them to decide how much effort a product page deserves, not to forecast your own rate.
| Measure | NRF figure |
|---|---|
| Holiday sales expected to be returned | 17% |
| Annual sales returned | 15.8% ($849.9 billion) |
| Online sales returned | 19.3% |
| Shoppers citing free returns as a major purchase factor | 82% |
| Retailers planning to extend return windows | 37% |
Four answers, one recording
Record one 40-second master per product in a single take on a phone: the item on a table, a hand measuring it, the item in use, the item being packed. The four short clips each answer one buyer question, so you can place them wherever the question gets asked.
Cut each with video trim: send start and duration, and leave precision at its default of exact so the cut is frame accurate. The master must already be a media.sume.com file in this workspace, because trim reads nothing else.
- Size: the item next to a ruler or a common object, 8 seconds.
- Material and care: a close shot of the surface and the care label, 8 seconds.
- In use: one real action, 8 seconds.
- What arrives: the box and everything in it, 8 seconds.
Join and caption
Timeline 1.0 takes one audio spine plus ordered video slots. For a silent set, declare audio.mode as silence with a duration_seconds of 32 and place the four trims back to back. Then burn the on-screen text with video captions using cues, which skips speech-to-text and works on silent clips.
Caption cues carry text, start and end in seconds, so the line "Fits a 15-inch laptop" sits over the size clip and nothing else.
Cost per product
The rates below are the flat public rates in the Sume docs. Timeline bills per output minute rounded up, and a caption job covers videos up to 60 seconds.
| Step | Jobs | Rate | Subtotal |
|---|---|---|---|
| Trim four answers | 4 | $0.02 per job | $0.08 |
| Timeline render, 32 s (1 minute) | 1 | $0.10 per output minute | $0.10 |
| Captions with cues | 1 | $0.20 per job | $0.20 |
| Total for the joined clip | 6 | $0.38 |
Where each clip goes
Place the size answer next to the size chart, the care answer beside the material description, and the packaging clip near the shipping and returns text. A shopper who is about to guess a size can see the answer where the guess happens.
In ads, use the joined 32-second version instead, since an ad has one slot and no page around it. The same four trims feed both, so the second use costs only the timeline render and the caption job.
What this does not do
A clip does not guarantee fewer returns, and NRF does not claim that. It makes the answer visible before purchase. Keep the four trims as separate files too, because the size answer can sit on the product page while the full set goes into an ad.
For a whole catalog, run the same recipe per product from a spreadsheet, and check a short batch of three before you commit the rest.
Sources
Related posts
More in Use cases
- Home Depot primary image: 1000x1000 white and what to ask Sume
Home Depot's content best-practices sheet: product only, plain white, 1,000x1,000 minimum. The Sume model, size and background settings that match.
- How many hook clips does $20 buy: Wan 3.0 vs Seedance 2.5
Twenty dollars buys 64 five-second Wan 3.0 clips at 480p. Whole-clip counts for Wan 3.0 and Seedance 2.5 at 5, 10 and 15 seconds, from the rate card.
- How to make a TikTok micro-drama with AI: six steps and the costs
A six step workflow for a vertical AI micro-drama episode with Sume: script, shots, first frames, assembly, captions and the is_aigc flag. About $52 for 90 s.
- How to remix a video with AI: four routes by what you change
Remixing a clip means a prompt edit, a person swap, a motion transfer, or a new clip from references. Which Sume route fits each, with the fields.
Written by Sume