Automating Product Ranking on E-Commerce Marketplaces

23. Brand Group operates multiple e-commerce brands across international markets. Its business involves launching and managing consumer products through online sales channels. 

Running products across different marketplaces created recurring work around product discovery, visibility, and marketplace engagement. Each storefront also introduced its own operating requirements. 

The project focused on making these recurring activities easier to manage through automation. This supported a more structured approach as product and marketplace operations expanded.

Product-ranking activities followed the same repeated process. An operator had to search by keyword, move through results, identify the correct listing, perform the required action, and repeat the workflow for additional products. 

The process became even more difficult when dealing with several marketplaces. This is because each storefront had different page structures, selectors, consent dialogs, pagination patterns, and account flows. Dynamic loading also interrupted browser interactions.  

For these reasons, the automation had to handle more than simple navigation. It needed accurate product matching, configurable inputs, stable browser sessions, and clear recovery paths when individual runs failed. 

Key Issues 

• Repeated product searches and marketplace actions required significant manual interaction.  
• Storefront layouts, navigation patterns, and account flows differ across platforms.  
• Product matching had to prevent actions from being performed on incorrect listings.  
• Dynamic pages and browser session issues interrupted automated execution.  
• Products, keywords, and run settings had to be consistent to keep things easy to update.  

An automation system was developed around marketplace-specific workflows and a shared utility layer. Google Sheets provided the operating inputs. Dedicated scripts handled marketplace interactions, while shared utilities managed browser behavior, pagination, proxies, and failure recovery.  

Spreadsheet-Driven Configuration 

Google Sheets was used to maintain product IDs, keywords, frequency settings, and delays. It ensured that routine changes could be made outside the codebase. This gave operators a simpler way to control each run without editing Python files.  

Marketplace-Specific Automation 

Dedicated Selenium workflows reproduced the actions required by individual storefronts. The verified implementations handled product search, result navigation, product matching, and cart or wishlist interactions. Product IDs were checked against listing URLs before any further action happened.  

Resilient Browser Sessions 

Undetected Chrome, Selenium Stealth, Selenium Wire, and proxy-enabled browsing were used to improve browser execution across regional storefronts. Explicit waits handled dynamically loaded elements while shared helpers supported pagination, smooth scrolling, and proxy configuration.  

Retry & Failure Handling 

Product cycles included retry logic when an interaction failed. Screenshots, logs, and failure records provided information for later investigation. This reduced the need to watch each browser session manually.  

Independent Marketplace Workflows 

Marketplace scripts were kept separate so each storefront could use its own selectors and interaction logic. Common functions remained in shared utilities. This reduced duplication while allowing platform-specific workflows to evolve independently.  

Key Deliverables 

• Developed marketplace-specific automation for product search, verification, cart, and wishlist actions.  
• Created Google Sheets controls for products, keywords, frequency settings, and delays.  
• Built shared utilities for browser setup, pagination, scrolling, waits, and proxy handling.  
• Implemented product-ID verification to improve listing accuracy before actions were performed.  
• Added retries, screenshots, logs, and failure reporting for unsuccessful runs.

More Consistent Product Identification 

Product-ID checks provided a defined verification step before cart or wishlist actions were performed. This reduced dependence on manual listing identification.  

Faster Routine Configuration 

Operators could change products, keywords, delays, and frequency settings through Google Sheets. New run requirements therefore did not require changes to the automation scripts.  

Clearer Failure Visibility 

Screenshots, logs, retries, and failure records made unsuccessful product cycles easier to identify and investigate.  

Easier Marketplace Expansion 

Shared browser utilities provided a reusable technical base for marketplace-specific scripts. New storefront logic could be added without rebuilding common automation functions.