Client
Relief Center is Russia's largest manufacturer of stationery and the leader in the wholesale distribution of goods for office, study and creativity. The portfolio includes 17 own brands and 6 production sites.
The situation at the start of the project
Relief Center has implemented the centralized Pimcore PIM system, which stores all trade names, specifications and media files — the total data volume is more than 30 TB. Information is regularly updated through integrations with 1C and ERP systems.
The company is actively working with Wildberries, Ozon and Yandex.Market marketplaces. Marketplaces account for a key share of sales; collections are updated regularly, and in order to publish new products, you have to manually adapt data to the format of each marketplace channel.
{{cta}}
Issues
- Manually bringing product data to the requirements of each marketplace — that is, “translating” attribute names and values from the PIM system — takes a lot of time for content managers and slows down the publication of products on external platforms.
- Updating collections requires considerable effort. Three full-time employees are exclusively dedicated to preparing and publishing Wildberries product cards for the company's leading brands, Berlingo and BG.
- Handwork also becomes a source of human error: inaccurate characteristics, missing fields, duplicates. This leads to additional costs and has a negative impact on brand reputation and sales.
- The growth in the range increases the load on the content department, limiting the speed of scaling.
Business result
In 2.5 months, we launched a service for comparing product category data in the PIM system with marketplace requirements, which allowed us to:
- Reduce the time it takes to launch new collections on the marketplace: updating items now takes 1 day instead of 10. Thanks to the service, one employee is able to download thousands of cards per day, compared to no more than 30.
- Free up to 100 work-hours per week: By automating the processes of preparing and uploading cards, the team can now focus on business tasks rather than manually publishing content.
- Avoid downtime if the marketplace changes the data reception format: you no longer need to manually update data for each product; it is enough to update compliance for a product category once.
Development process
Business request at the start
Set up the automatic publication of product cards from the PIM system to marketplaces.
The goal of the project at the start
Create a service for intelligent attribute comparison between the client's internal PIM system and marketplace directories. The first step is to launch Wildberries as a key sales channel.
What was it like before the introduction of the service
The process of preparing and publishing product cards for marketplaces takes considerable time and requires the high involvement of the content team:
- Each marketplace uses its own data formats: field names, directories and values differ, and data from the client's PIM system must be manually adapted to the format of each marketplace.
- Adaptation includes not only renaming fields, but also comparing values (for example, colors, materials, sizes, categories, etc.) with each site's directories.
- The work is carried out in separate personal accounts of each marketplace. Content managers have to publish cards, track statuses, fix errors, and update data — for each marketplace separately.
- One employee can publish an average of 20—30 cards per day, but mistakes and corrections constantly occur: a discrepancy between formats or values leads to publications being rejected and then adjusted.
Development time
2.5 months — from the start of the project to the start of operation by business users.
How the service works
The matching service is installed on the client's infrastructure and connected to its PIM system. It is designed to automate the publication of product cards on marketplaces and manage attribute comparison.
The service performs several key functions at once:
- Centralized management of categories and directories: all attribute comparison logic has been placed in a separate module that can be configured without interfering with the main PIM system. This ensures the security and flexibility of the architecture.
- Category matching interface: through the web interface, the user can select a marketplace category, configure attribute matching, and start publishing. If the selected category is not configured, the service notifies you when you need to set up notifications.
- Data comparison: KT.mapper allows you to adjust the correspondence between fields and values in the client's PIM system and the requirements of each marketplace. This applies to both attribute names (e.g. “color”, “brand”, “material”) as well as the values themselves (e.g. “Green”, “Green”, or color code).
- Integration with marketplaces: after setting up the comparison, the service automatically converts product cards into the format of the desired marketplace and publishes them directly, without the need for manual work in personal accounts. The service receives publication statuses from marketplaces and transfers them back to PIM, allowing you to track the entire card life cycle.
The advantages of the chosen solution
- The service can be easily scaled to other marketplaces without improving the internal logic of the PIM system.
- The service adapts data from the PIM system to the required formats, but cannot affect the original data.
- The service can be easily transferred to another PIM system and its functionality expanded.
Technology stack
Interface: Pimcore
Backend: php, Symphony, Kafka
Future plans
Right after the launch of the service on Wildberries, we started preparing integrations with Ozon and Yandex.Market. Within a month from the date of our release on Wildberries, we plan to launch publications on these sites as well.
In the future, we will consider integrating AI models to fully automate the process, as well as adding video support — currently only texts, photos and reference materials are uploaded.
{{cta}}
