Shoupapp
Journal

How to Automate Product Photo Editing

Automate product photo editing: which steps to automate, how to turn decisions into settings, pick a trigger, test the workflow and keep it reliable.

Most product photo editing is the same work done again: cut out, add a shadow, resize for each channel, write a description, repeat with the next batch. Tools that apply one edit to many images help, but they still wait for someone to press the button. To automate product photo editing properly, the process has to be written down, every decision has to become a setting, and something other than a person has to start the run. This guide walks through that, with a worked example, and covers what keeps an automation reliable once nobody is watching it.

An automation is a written-down process

Before any tool, write down what happens to a new product photo today, step by step. Then, next to each step, write the decision someone makes and the setting that would replace it.

StepDecision todaySetting that replaces it
Remove the backgroundWhich background for which channel?White for the marketplace, brand color for the store
Position the productHow much space around it?Centered in a fixed safe area, same margin for every photo
Add a shadowWhich style?Contact shadow for every product
ResizeWhat size for each channel?2,000 × 2,000 px JPG for the store, a white-background version for the marketplace
Write the descriptionWhich tone and length?Professional tone, standard length, per channel

If a step has no clear setting, the process is not ready to automate yet. Decide the rule first.

Choose what starts the run

TriggerBest for
New uploadEvery new photo processed as soon as it arrives
ScheduleA daily, weekly or monthly run, such as new stock overnight or content every Monday
A list or CSV fileOne-off jobs, such as fixing a list of old, low-quality photos

When products need different treatment

A single catalog often mixes products that should not be treated the same: shoes on a contact shadow, jewellery with a larger margin, apparel sent to virtual try-on. Instead of separate manual passes, use conditions: products of one type go down one branch, the rest down another. The rule stays one workflow, and nobody has to sort the photos first.

Test before switching it on

Run the workflow once on a small set that represents the catalog:

  • An easy product: a box or a bottle on a plain background.
  • A hard edge: hair, fur, mesh or thin straps.
  • A very small or soft supplier photo.
  • A transparent or reflective product.
  • One product from each branch of your conditions.

Check every result against each channel's requirements. Adjust, test again, and only then switch the rule on.

What keeps an automation reliable

  • A report after every run. A problem noticed the same day costs a re-run. One noticed a week later costs a week of wrong images.
  • Failed steps handled, not hidden. A photo that fails should be reported, and you should not pay for work that did not complete.
  • No reprocessing of its own output. A scheduled rule that picks up the images it created last night will process them again, and again. It should only take what is genuinely new.
  • A review of new product types the first time they go through.

What should stay manual

Automation is for decisions you have already made. Keep people on the new ones: hero campaign creative, the first run of a new product category, and anything published to a product page for the first time.

The cost logic

An automation should cost the same as doing the steps by hand. What it saves is the time spent preparing, starting and checking each batch, and the inconsistency that creeps in when different people run the same process slightly differently.

Three workflows that pay off first

  1. New stock, ready by morning. Photos uploaded during the day are cut out, given a shadow, resized for each channel and described overnight.
  2. Weekly content. Every Monday, last week's new products become banners or short videos.
  3. Catalog clean-up. A CSV list of old photos is upscaled or unblurred.

Agencies can keep one rule per client, each with that client's settings and schedule.

Building it on Shoupapp's canvas

Shoupapp's automation is a visual canvas: you drag the steps in and connect them, with no code. A rule starts on a new upload, on a schedule (daily, weekly or monthly, at the hour you choose) or from a CSV file. The steps are the same ones you use by hand: background removal, shadow, crop, padding, resize, enhancement, logo, AI edit, upscale, unblur, new angles, try-on, banners, descriptions and video. Conditions, waits and email steps handle the logic.

You can run a rule once on a photo of your choice to check it before switching it on. After every run you get an email; a step that fails returns its credits; and a rule never reprocesses its own output. Up to 1,500 photos can go through a single run. Each step costs what it costs by hand, and the rule itself is free.

Where results go

An automation that produces hundreds of files a week needs a plan for them. Decide in advance where each output lands and how it is named: one version per channel, kept next to the original, with the product's SKU in the name if you export to other systems. The original should never be overwritten, so a rule can always be changed and run again from the source.

From one rule to a system

Start with one rule for the most repetitive task and let it run for a few weeks. Once it is reliable, add the next step to the same rule or build a second rule for a different trigger. Growing one working rule is safer than launching five at once, because when something goes wrong you know which change caused it.

Setting it up for an agency

Agencies handling several clients benefit most from keeping one rule per client. Each rule carries that client's background, margin, shadow and export sizes, runs on that client's schedule and sends its report to the person responsible for the account. A new shoot from a client then goes through the same treatment as the last one, whoever is on shift.

Automation questions

Do I need to know how to code?

No. The steps are dragged onto a canvas and connected.

What happens if a step fails?

Its credits are returned, and the email after the run tells you what happened.

Can I pause a rule?

Yes. Switch it off, change the settings, test again and switch it back on.

Start with the one task your team repeats every week without thinking, usually cut-out and resize for new stock. See how to remove the background from multiple images for the settings that task needs.

Send us five photos.

We will run them the way we would run your whole catalog and send back exactly what came out — before you decide anything.