Planning an AI content budget for e-commerce brands
Product photography and video scale with your catalog, not your calendar. A 200-SKU launch costs differently than a 12-SKU restock, and most subscription pricing doesn't tell you which one you're planning for.
Budget per SKU, not per month
A monthly subscription plan is built around a person's workflow: one designer, a steady pace of edits, maybe a few dozen assets a week. E-commerce catalogs don't move that way. A new product line can mean 40 SKUs landing at once, each needing 3-6 angles, a lifestyle shot, and a couple of size variants. Estimating "how much AI image generation costs this month" is the wrong question; "how much does this SKU batch cost, at this shot count per SKU" is the one that maps to your actual work.
Multiply shots-per-SKU by cost-per-image, then add a retry allowance (see the next section) before comparing that number to a flat-rate plan. For steady, predictable volume a subscription can still win. For lumpy, launch-driven volume, usage-based pricing on a provider like Black Forest Labs (FLUX) often tracks your actual spend more honestly than a seat-based plan sized for someone else's workflow.
Where retries eat the budget
The published cost-per-image number is never the number you actually pay per usable image. Product shots have a harder bar than general content: incorrect logos, warped text on packaging, an extra strap or button that doesn't exist on the real product, and lighting that doesn't match your brand's other photos are all common failure modes that don't show up as clearly in casual test prompts.
Budget a real retry multiplier for product work — often 1.5-3x the raw image count for anything with fine detail (labels, logos, small hardware), and closer to 1x for simple lifestyle or background-only shots. If you haven't measured your own retry rate yet, run a small batch of 10-20 real SKUs first and use that ratio going forward rather than trusting a vendor's best-case demo numbers.
Keeping a product consistent across shots
The other cost that's easy to miss: consistency work. Getting the same product to look identical across a hero shot, three lifestyle angles, and a size-comparison graphic often takes more prompt iteration and reference-image tooling than generating any single image alone. Some workflows lean on image-to-image or reference-conditioned generation to hold a product's shape and color constant, which can mean paying for a first "anchor" generation plus several conditioned variants rather than treating each shot as an independent cost.
If your catalog has a few hero products that appear across many pages (a bestseller, a seasonal item), it's often worth budgeting those separately at a higher per-shot cost, since they'll need more iteration to get exactly right, while treating long-tail SKUs at a lower, bulkier rate.
Generated images vs. a photo shoot
Generated product photography isn't a universal replacement for a physical shoot — it's strongest for background variations, lifestyle scenes, and quick seasonal refreshes of an existing product photo, and weaker when a buyer needs to trust an exact, unretouched view of texture, fit, or scale (some marketplaces and categories also have their own rules about generated vs. photographed listing images, worth checking before you commit a catalog to one approach). A common working pattern is a single real photo shoot for the anchor product images a buyer relies on, with AI generation used for the higher-volume, lower-trust surfaces: ad creative, seasonal banners, social variants, and background changes.
Costed that way, the AI portion of the budget covers volume and iteration speed, not the core trust-building shots — which keeps the comparison honest instead of pitting a full shoot against a full generation run as if they served the same purpose.
Frequently asked questions
How many images should I budget per SKU?
It depends on the category, but a common baseline is one hero shot, 2-4 lifestyle or angle variants, and one size or detail callout — call it 4-6 raw generations per SKU before retries. Apparel and anything with fine print or logos tends to need more; simple household goods often need fewer.
Is a usage-based provider always cheaper than a subscription for catalog work?
Not always — it depends on how steady your volume is. A subscription can be cheaper per image at high, consistent volume; usage-based pricing is usually a better fit when volume is lumpy (launches, seasonal drops) since you're not paying for idle capacity between batches.
Does this cover video as well as images?
The same per-unit budgeting logic applies to product video (unboxing-style clips, 360 turntables, short ad cuts), but video costs scale with both duration and resolution on top of retry rate, so it's worth estimating video and image batches separately rather than blending them into one number.
Sources and corrections
Provider pricing referenced here is drawn from the tools listed in our tools directory, cross-checked against each provider's own pricing page. Methodology for how we calculate per-unit costs is documented on our methodology page. If you spot a figure that's out of date, please let us know the page and inputs so we can correct it.