Guide 05
E-commerce · Updated 2026-09-29
Batch-Processing Product Photos for Your Online Store
Product catalogs die by a thousand inconsistent photos: some 4000px, some 800px, mixed formats, nobody knows why the page loads slowly. Here's a repeatable pipeline — one width, one quality, one format — with the arithmetic that explains why it works.
The Pipeline
- Fix the delivery width: 1000–1200px covers every mainstream store's zoom view with retina headroom. Measured cost at 1200px, Q70: ≈104 KB per photo (vs ≈191 KB at 1600px).
- Resize everything to that width — the Resize tool uses the same high-quality browser downscaling we measured in our resizing study (0.54× file size going 1600→1200, fidelity preserved).
- Batch-compress at one quality: drop the whole folder into the compression tool — it queues multiple files — and set quality 70. Consistent settings make a catalog look professional; per-file improvisation makes it look like a yard sale.
- Format: WebP if your platform accepts it (measured ~31% smaller than JPEG at matched fidelity for photos — see the format study); JPEG otherwise.
The Arithmetic That Pays for Itself
Take 100 product photos from a modern phone/camera (≈4000px, ≈3 MB each):
| State | Per photo | Catalog (×100) |
|---|---|---|
| As shot from camera | ≈3,000 KB | ≈300 MB |
| Resized 1200px + Q70 | ≈104 KB | ≈10 MB |
A 30× reduction, and the 1200px versions are the ones that actually match what shoppers see. Category pages that loaded 20 thumbnails now pay a fifth of a single original photo.
Two Extras Worth 5 Minutes
- EXIF: re-encoding strips it (measured in our EXIF study) — so your studio's GPS coordinates don't ship with every product photo. If your platform takes original uploads, check one photo with the Image Info tool first.
- Work locally, upload nothing: a 300 MB catalog batch is exactly the kind of payload upload-based tools choke on (and charge for). Browser-side processing runs at your CPU's speed and the photos never leave your machine — the architecture is measured in our network comparison study.