Capture once
One real reference packshot per SKU is the only physical step. Everything downstream is graph work.
Ecommerce product photography at catalog scale: the studio math that finally favors the canvas
"Ecommerce product photography" and its mirror phrase "product photography for ecommerce" together pull ~1,070 US searches a month; "AI product photography" adds ~390 at difficulty 11. The search volume is small; the cost problem it names is enormous. 2026 US studio rate cards cluster tightly: $25–$50 per white-background packshot (median ~$33 across 22 published cards), $100–$500+ per styled lifestyle or on-model image, $800–$2,500 for a production day. A 200-SKU launch needing six images per product is a $30,000–$60,000 studio line and an 8-week sample-shipping calendar. The canvas inverts that: one real reference per SKU, then every channel's set derived from a saved graph — while the marketplace specs that actually gate your listings (Shopee, Amazon, Shopify, Meta, TikTok) stay enforced per export preset.
Reviewed October 2026 · Prepared by Berrys · Technical specifications & Canvas node workflows
The cross-channel spec matrix (all checked 22 Oct 2026)
| Parameter / Slot | Marketplace & Production Requirement |
|---|---|
| Shopee | 1:1 · cover on solid bg, product ≥60% of frame, no text/watermarks, logo top-left ≤10% · 9 slots · ≤2 MB |
| Amazon | Pure white RGB 255×3 main, product ≥85% of frame · 1,000+ px long side for zoom · MAIN shows the product once, no props · 7+ slots + video |
| Shopify | 2048×2048 master, one ratio per catalog, zoom floor 800×800, ≤20 MB |
| Meta ads | 4:5 1440×1800 feed · 9:16 1440×2560 Stories/Reels · 14/35/6 safe zone · ≤30 MB img / ≤4 GB vid |
| TikTok | 9:16 video 1080×1920 · safe zone 160/440/80 px + right rail · 5–60 s policy, 21–34 s wins · ≤500 MB |
| Google Shopping | ≥100×100 (250×250 apparel), 800×800+ recommended · real product on clean bg · no watermarks/logo overlays on apparel |
| Cost baseline | US studio: $25–$50/packshot, $100–$500+/lifestyle, $800–$2,500/day · AI-assisted agency work: $40–$200/image · Canvas: credits per render, human time only at the packshot |
Free Browser Tool
Run free audit →Check your cover photo against Ecommerce content teams rules
Test your packshot cover image for solid white background (#FFFFFF), product frame-fill percentage and 1:1 aspect ratio before uploading to Seller Centre.
The creative-refresh pipeline
Channel presets
Export nodes carry per-platform rules — Shopee ≤2 MB squares, Amazon 85%-fill white, Shopify 2048 masters, Meta 4:5 + fan, TikTok 9:16 — one run produces the whole matrix.
Listing copy pairing
Alt text and image-slot naming come out of the graph too (slot 4 macro → 'close-up of…'), which is where ecommerce image SEO actually lives.
Refresh loop
Ad-fatigue variants, seasonal scenes and new-channel crops derive from the same lock — the catalog gets richer between shoots instead of before them.
The three failure modes that cost money, not clicks
Thumbnail click math
On Amazon and Shopee your image is a 120–200 px thumbnail in a row of 20 competitors. Frame-fill ratio (85%/60% rules exist because of this) beats prettiness: a beautiful scene where the product fills half the frame loses the click to an ugly-but-large product. The canvas exports cover-cropped masters at the fill ratio your category won, measured, not guessed.
Cross-channel ratio stress
Teams that re-crop per channel by hand end up with five different products: shifted perspective, clipped detail, color drift per channel. Branch crops from one master preserve geometry; overlay text re-composes instead of pixels getting sliced.
The color-accuracy cliff
Screen-to-eyes color error is the top 'not as described' return driver for home, beauty-adjacent and apparel SKUs. Batch generations inherit the same lighting profile, and the proof workflow (render vs physical swatch) is part of the graph, not an afterthought.
How Berrys runs the catalog
Per-SKU references at catalog scale
500 SKUs stay individually locked — geometry, label typography, variant colors — because batch runs fan from each SKU's own File node reference, not from prompt memory.
Agency + platform in one
Berrys is a self-serve canvas (Berrys Canvas) and a done-for-you studio (Berrys Studio, ex-Brainchild art direction). Teams run their own graphs; brands can hand the brief to the studio and get finished channel sets back.
Credit math you can audit
One catalog run's credit cost is measurable against the $25–$75-per-image studio line — the spreadsheet every ops lead actually needs before approving tooling.
Production workflows in Berrys Canvas
One upload → 6-channel omnichannel matrix
1. File Node: Upload raw product packshot.
2. Brand System Node: Attach master brand profile and multichannel compliance rules.
3. Image Node (Nano Banana 2.1): Render 2048×2048 master hero.
4. Multi-channel fan-out: Automatically export Shopee 1:1 (≤2MB), Amazon 85% pure white, Shopify 2048px, Meta 4:5, and Google Shopping clean-background assets.
200-SKU catalog launch compression
1. File Node: Ingest multi-product catalog via List Node.
2. Brand System Node: Enforce locked camera angle, lighting temperature, and contact shadow physics across every SKU.
3. Image Node (Nano Banana 2.1): Batch render all products into release-ready gallery folders in one automated pass.
Ad-to-PDP visual continuity sync
1. File Node: Select top-converting lifestyle advertising scene.
2. Brand System Node: Bridge creative directly to marketplace A+ content and Shopify secondary slots.
3. Image Node (Flux 3): Export wide-format banner crops keeping product pixels 100% identical.
Berrys Plugins · Model Context Protocol (MCP)
Create in chat. Keep it on your canvas.
Already working in an AI agent? Connect Berrys to Claude, Cursor, ChatGPT, Codex, or Hermes. Your agent plans the ecommerce content teams suite and runs the shots on your Berrys canvas, with saved Brand Memory and credit approval in chat.
Connector URL:
https://app.berrys.io/api/mcpRemote MCPDrop-in Agent Brief (Ecommerce content teams)
Using Berrys, generate an omnichannel 6-channel e-commerce asset suite from my product reference in ./assets: 1. File node: Ingest master product packshot. 2. Brand System node: Set universal studio lighting profile. 3. Image node (Nano Banana 2.1): Render 2048×2048 master hero packshot. 4. Channel fan-out: Export Amazon pure white (#FFFFFF, ≥85% fill), Shopee 1:1 square (≥60% fill, ≤2MB), Shopify 2048px master, and Meta 4:5 ad crop. Estimate credit cost and request approval before generating.
Compliance & production boundaries
This is a scaling argument, not a truth argument. The canvas only looks after product geometry because the input is a real reference — a catalog with no genuine photography to seed it will fabricate beautifully and return catastrophically. Keep real capture for hero launches, fit/texture-critical SKUs and any gated marketplace slot; use the graph to make one real input answer 40 channel questions.
When NOT to use AI generation for this
- Products that don't physically exist yet (renders become contractual promises)
- Marketplace main-image slots where the platform requires a true photograph
- Scale-critical goods (rugs, furniture, tools) without real room/size references
Frequently asked questions
How much does ecommerce product photography actually cost?
On 2026 US rate cards: $25–$50 per white-background packshot (median ~$33), $100–$500+ per styled lifestyle image, $800–$2,500 per production day. Catalog-scale teams pay per-SKU × per-image × re-shoot frequency — which is the line the canvas collapses to credits + one reference photo.
Can AI images rank on Google Shopping?
Google's policy is about accuracy and identity (real product, clean background, no watermarks), not about how the pixels were made. Derive renders from real packshots and they behave like any product image.
What about marketplaces that ban AI main images?
Where a platform requires an actual photograph of the item (some category-specific Amazon/Grocery lines, and any 'show the real product' rule), the workflow keeps the real photo in the gated slot and uses generated imagery only where allowed.
Does this replace our retoucher?
It replaces per-image retouching volume ($0.70–$3.00/image is the common outsourcing rate). Your QC person stays — they move from cloning dust to reviewing batches.
Related tool comparisons & guides
In-Depth Guides & Research from Berrys Journal
Create on-brand assets without booking a studio
Start free on Berrys Canvas — 50 credits, no card required — or send Berrys Studio a brief and get finished campaign visuals with agency art direction.