Client Context
An automotive parts marketplace with 500,000 SKUs across three channels.
The client operates a large-scale automotive parts marketplace, selling across Amazon, eBay, and Shopify. Their catalogue of approximately 500,000 SKUs spans parts from hundreds of manufacturers — each arriving as raw, inconsistent supplier data that needs to be transformed into polished, channel-ready listings.
Growth was constrained not by demand, but by the throughput of their catalogue operation. Every new SKU required the same manual effort regardless of volume.
The Challenge
A catalogue operation bottlenecked by manual listing creation.
An automotive parts marketplace managing approximately 500,000 SKUs across Amazon, eBay, and Shopify was drowning in manual data work.
Each product required roughly 20 minutes of human effort to turn messy manufacturer data into polished multi-channel listings. The full processing pipeline ran on a 9-day cycle, slowing how quickly inventory could get to market.
The real cost was not just labour. It was the drag created by slow catalogue turnover across three channels that each required tailored output.
500,000 SKUs to enrich
Each product required titles, descriptions, bullet points, and metadata for three separate marketplaces — all from messy manufacturer data.
20 minutes per product
Manual enrichment of a single SKU took roughly 20 minutes of human effort, making catalogue-wide processing economically unviable.
9-day processing cycle
The full pipeline ran on a 9-day cycle, creating a structural lag between inventory arrival and the point at which it could reach customers.
Three channels, three formats
Amazon, eBay, and Shopify each required tailored output, multiplying the manual effort rather than allowing any shared enrichment logic.
What We Built
A rebuilt data and content pipeline for channel-specific listing generation.
60x rebuilt the end-to-end enrichment and content generation workflow so titles, descriptions, bullet points, meta tags, and marketplace categorisations generate simultaneously across all three sales channels.
Data enrichment
Ambiguous manufacturer data is resolved automatically, including incomplete codes and missing product names.
Channel-specific content
Amazon, eBay, and Shopify outputs are generated in parallel rather than one listing at a time.
Low-touch review
Approximately 97% of items now require no human review before publication.
Instead of treating content generation as a downstream manual task, the new system made structured enrichment and publishing logic part of the same automated pipeline.
The Results
From a 9-day bottleneck to a few hours of throughput.
Case Study · E-Commerce & Automotive
- Per-SKU processing: from about 20 minutes of manual effort to 1 to 3 seconds.
- Full pipeline cycle: compressed from 9 days to a few hours.
- Coverage and accuracy: rich content now publishes across three channels with 97% requiring zero human review.
Compressing the pipeline from 9 days to a few hours did more than save staff time. It removed the operational bottleneck that limited how quickly 500,000 products could reach market.
The business can now onboard new product lines and respond to demand shifts at a speed that was previously impossible with manual enrichment.
For a catalogue business at this scale, that speed becomes a structural advantage rather than a back-office efficiency win.