Case Study - Automating Product Image Processing for Multi-Retailer Ecommerce

The Challenge

A UK-based sustainable womenswear brand works with multiple retail partners including major department stores and online marketplaces. Each retailer has strict requirements for product image naming conventions—and they're all completely different.

The brand's internal system names product images using the size 10 SKU followed by a sequence number (e.g., ABC123_1, ABC123_2, ABC123_3). But each retail partner requires something different:

Department Store A requires the first image to use just the SKU with no suffix, while subsequent images use _2, _3, etc.

Online Marketplace B requires matching each SKU to a Consumer Code from their provided spreadsheet, then renaming all images using that code with sequential suffixes.

International Retailer C has the most complex requirements—matching SKUs to Product Numbers from a spreadsheet, then applying different suffix patterns based on image type: _M1, _M2, _M3 for model shots, _F for flat or cut-out images, and _L1, _L2 for lifestyle photography.

Previously, the team was manually renaming thousands of images for each retail partner. This process was:

• Time-consuming: Hours of manual work per product range, repeated for each retailer
• Error-prone: Incorrect naming led to rejected uploads and costly delays
• Complex: Staff needed to remember different rules and cross-reference spreadsheets manually

They needed an automated solution that could handle each retailer's specific requirements while being simple enough for their team to use.

The Solution

We built Image Processor Pro, a custom desktop application using Electron and Next.js. The app provides an intuitive wizard-based interface that guides users through the image processing workflow:

1. Select retailer — Choose from supported retail partners
2. Configure and select files — Pick the images to process
3. Process with real-time feedback — Watch progress and review results

Key Features

Intelligent Spreadsheet Matching

For retailers that require code mapping, users simply upload the retailer's spreadsheet. The app automatically matches each image's SKU to the corresponding Consumer Code or Product Number, then applies the correct naming convention. No manual lookups required.

Image Type Detection

For retailers with complex suffix requirements, the app supports different naming patterns based on image type—distinguishing between model shots, flat/cut-out images, and lifestyle photography with the appropriate suffixes for each.

Batch Processing

The app handles hundreds of images simultaneously, with real-time progress tracking showing exactly which files are being processed, their original names, and their new retailer-compliant names.

Safe Operations

Files are copied rather than modified—originals are never touched. The output folder opens automatically when processing completes, ready for upload to the retailer's system.

License Server

For distribution to the client's team, we implemented a serverless license validation system using Vercel and Redis, ensuring only authorised users can access the application.

Technical Implementation

The application architecture demonstrates modern desktop development practices:

• Electron 38 for cross-platform desktop capabilities
• Next.js 15 with React 19 for the modern, responsive UI
• TypeScript throughout for type safety
• Modular processor architecture — each retailer has a dedicated processor module
• 63 comprehensive unit tests covering all business logic

A significant refactoring effort reduced the main component from 760 lines to 160 lines—a 78% complexity reduction—while maintaining full functionality and improving maintainability.

The Results

The Image Processor Pro has transformed the client's workflow:

• Hours reduced to minutes — What previously took a full day of manual renaming now completes in under a minute
• Eliminated upload rejections — Automated validation ensures every file meets retailer specifications before processing
• 3 retailers supported at launch — Each with completely different naming requirements handled seamlessly
• Extensible architecture — Adding new retail partners requires only a new processor module following established patterns

The client's team now processes thousands of product images across all retail partners in a fraction of the time, with confidence that every file will be accepted on first upload.

What we did