Desport Case (ex‑Decathlon)
How an Enterprise-Level Reseller Scaled His Business with WB API
Introducing Desport
Desport (formerly known as Decathlon Russia) is one of the largest sports goods retailers in Russia. Seeking to expand their reach and improve sales efficiency, the company decided to enter several marketplaces. Among them, Wildberries — a Russian platform with a massive customer base and significant potential for scaling sales — played a key role.
We spoke with Ilya Botvinnikov from Desport, who shared how a major international retailer managed to scale sales through Wildberries without increasing the team size and while maintaining profitability.
«We increased turnover sixfold in the first year, cut order processing costs by nearly 40%, and automated almost everything — from receiving orders to accounting. Now our daily volumes are comparable to flagship stores, and we’ve achieved all this without expanding our staff»
— Ilya Botvinnikov, Desport
How did the company achieve such results? What solutions fueled this growth? And what can other Wildberries sellers learn from this experience?
In this article, we’ll take a detailed look at how the seller solved numerous challenges through the smart implementation of API integration, process automation, a customized employee role model, and efficient stock and pricing management. This case will be useful to Enterprise-level companies as well as any other sellers planning to scale their sales via Wildberries.
Summary
Desport (formerly Decathlon Russia) successfully launched on Wildberries, overcoming key scaling challenges through automation and API integration.
Main objectives:
- Automate the processing of thousands of daily orders without increasing headcount.
- Manage inventory and pricing in real-time to minimize errors and margin loss.
- Adapt a large product assortment (15,000 SKUs) to meet Wildberries requirements.
- Automate accounting processes and returns management.
Implemented solutions:
- Developed an «order orchestrator» integrating WB API with SAP ERP via Apache Kafka, enabling processing of up to 30,000 orders per day.
- Customized employee role model, including a night shift with minimally qualified staff, significantly reducing costs and errors.
- Introduced a two-level order picking control to minimize mispicks and losses.
- Created an inventory «gap» management mechanism to eliminate cancellations due to stock-outs.
- Built a specialized PIM system to automatically adapt product cards to Wildberries specifications.
- Implemented nightly automated generation of accounting reports, with plans for full integration into financial systems.
Results within one year:
- Sales turnover on Wildberries increased 6-fold.
- Operational staff reduced by 40%.
- 80% of orders came from regions outside existing offline store coverage.
- Achieved near-total process automation, eliminating manual tasks.
Conclusion — for large businesses, process automation through WB API integration is critical
Challenges and Objectives
The main goal for Decathlon Russia when joining Wildberries was to scale sales rapidly and effectively, at the level of a major international chain. Early on, the company faced difficulties common to most Enterprise-scale retailers venturing into marketplaces:
How can we automate handling thousands of daily orders, maintain control over a vast product range, and eliminate human error in key processes?
Decathlon Russia began by processing orders manually, using Google Sheets and custom Python scripts. However, this approach became critically inefficient once orders grew to several thousand per day, and during major sales events, it was impossible to manually handle more than 30,000 orders in a day.
«In the beginning, we only had a manual workflow using Google Sheets and Python scripts. We realized this model could lead to increased costs, penalties, and errors at each step of order processing. It was clear that without full automation and a proper API, we wouldn’t get far»
Large Number of SKUs
Another challenge was the retailer’s wide product assortment — around 15,000 SKUs. Each product required frequent updates on stock levels, pricing, and order status. Manually updating this data wasn’t just time-consuming; it was physically impossible. This severely limited growth potential and led to constant data mismatches between the warehouse and marketplace.
Problems with Manual Processing
Manual order processing always left room for human error. With such a large volume, even a small percentage of mistakes could have serious consequences: from marketplace penalties to a drop in store ratings and negative customer feedback. Additionally, mispicks (sending the wrong product) occasionally occurred.
Stock and Pricing Management
The company sells through multiple channels at once: its own brick-and-mortar stores, an online store, B2B clients, and several major marketplaces, including Wildberries. Each channel draws on the same product inventory, requiring near-instant updates on stock levels.
Without timely and automated data synchronization, the company risked:
- Selling out-of-stock products.
- Canceling orders and receiving negative customer feedback.
- Lower marketplace ratings and financial penalties.
«If the warehouse has, for instance, only 100 units of a product while sales are happening simultaneously on Wildberries and other channels, we must constantly keep stock data up to date. Without an API and automated data exchange, maintaining that accuracy is practically impossible. An error in stock levels can be very costly, especially for high-value items like bicycles or professional sports equipment»
The «Gap» Issue: Avoiding Out-of-Stock and Order Cancellations
To reduce the risk of selling unavailable items, the seller was forced to artificially understate the reported stock, showing the marketplace fewer units than they actually had.
This approach addresses two important factors:
- A buffer against sudden surges in demand.
- Protection from risks posed by large orders coming from other channels (e.g., unexpected B2B bulk orders).
Without automated gap management, it was nearly impossible to manually keep real stock data accurate. Any error could lead to cancellations.
Safeguarding Margins Under Auto-Discount Conditions
Auto-discounts posed another significant challenge. These are discounts that the platform automatically applies to certain products.
«This was a serious problem. We calculated each product’s margin in advance, factoring in its cost and the marketplace commission. But a product could get an automatic discount. Without a mechanism to quickly track these discounts, our margins dropped sharply. We needed a way to react immediately and remove the product from sale if necessary»
Technical Constraints
How do we build a unified technical infrastructure that can consolidate orders, stock levels, pricing, and product information while ensuring reliable WB API integration?
High-Load Constraints of SAP ERP’s REST API
Decathlon employed SAP ERP as the central tool for managing orders, inventory, and pricing globally. Initially, data exchange with SAP took place via a REST API. However, with the sudden increase in orders that needed transferring between Wildberries and SAP, the company hit critical load limits. This triggered the need to explore an architecture capable of handling large volumes of data without losses or delays in order processing.
«Our ERP system’s REST API couldn’t handle the load once we had thousands of orders daily. During peak sales, the volume of requests was so high that we practically ‘brought down’ SAP’s interface. It became evident that we needed a completely different technical solution for these data volumes»
Fragmented Systems for Orders, Stock, Pricing, and Product Pages
Another hurdle was the fragmented internal IT landscape. Order management was in one system, stock and pricing in another, and product information (catalog data) in yet another. These systems were neither directly integrated with each other nor had direct integration with Wildberries. Without a single management hub, delays and data errors were unavoidable.
Content Specifics for Marketplaces and PIM
How can we effectively manage thousands of product listings, adapting them to the platform’s requirements without constant manual labor and mistakes?
The Need for a Separate Mechanism to Manage Product Listings
Selling through marketplaces requires a special approach to product content. Even though Wildberries has relatively lenient requirements for descriptions, images, sizes, and colors, these often differ from the retailer’s internal catalog standards.
«Imagine we label a product as ‘dark blue,’ but Wildberries needs it to say just ‘blue’ or ‘light blue.’ If we don’t meet the marketplace requirements, the product listing won’t pass moderation. We needed a tool that could automatically align our data with the platform’s rules»
Decathlon initially used its own Product Information Management (PIM) system to manage product data. However, the company’s standard PIM was not at all suited to Wildberries and other major platforms. Many required fields were simply missing, so each listing had to be reworked manually, consuming massive time and resources.
«At some point we realized our standard PIM solution wasn’t suitable for Wildberries. We had to create a separate, specialized PIM for marketplaces that added all the essential fields and attributes we’d never needed before. Without this dedicated tool, we were spending too much time on manual work, seriously delaying product launches on the platform»
Accounting Reports and Returns
How can we ensure automated data exchange with accounting and create a reliable order identification mechanism to minimize manual labor, reduce mistakes, and optimize financial tracking?
Another serious issue for Decathlon was automating accounting and processing returns. As sales through Wildberries soared, the number of transactions that the accounting department had to handle manually also surged. This led to multiple critical problems that directly impacted the company’s operational speed and financial accuracy.
Difficulties with Automated Returns Handling and Reporting
Before automation, the accounting team had to manually download sales and return data from the Wildberries seller portal. This caused delays, required substantial labor, and increased the risk of human error.
«Our accounting team was forced to manually download dozens of reports every day just to keep track of incoming data. These reports often mixed FBS and FBW data, and accountants couldn’t determine which items were returned to our warehouse and which stayed at Wildberries. This made mistakes more likely and reconciliation very difficult»
Order ID Problems (Seller’s Internal Number vs Wildberries Order ID)
Yet another major challenge was matching orders to their identifiers. Sellers often have internal order numbering systems that differ from marketplace-generated IDs. Wildberries sales and return reports include only the marketplace order ID, without the seller’s internal number. This complicated reconciliation and order identification.
«If a Wildberries order sticker was lost, reconnecting it to our internal order number became nearly impossible. Our accounting department couldn’t properly record the return, so they had to search manually and cross-reference data. It was very time-consuming»
Solutions
Creating a Unified «Order Orchestrator»
Realizing that successful scaling depended on full process automation, Decathlon decided to develop its own integration system modeled as an «order orchestrator» This system became the central hub for all order, shipping, and return operations and significantly streamlined the company’s interaction with Wildberries.
Integrating with WB API for Orders, Shipments, and Returns
The first step was establishing a stable integration with the WB API. The system began automatically receiving orders, updating statuses, and passing shipment and return information in real time. Orders came directly from Wildberries into the orchestrator, then automatically distributed to warehouse systems and the company’s ERP.
«Once we launched comprehensive API integration, we gained the ability to process tens of thousands of orders daily with no manual intervention. This completely eliminated human error and provided stable, transparent warehouse operations»
Switching to Kafka and Accelerating Interaction with SAP ERP
Because the initial SAP REST API approach wasn’t up to handling peak loads, the company moved to a high-performance architecture based on Apache Kafka for data exchange between the orchestrator and SAP ERP. This approach dramatically sped up data transmission, reduced the load on ERP, and ensured no order loss or delays even at extreme volumes.
«Kafka solved our load issue completely. Even with over 30,000 daily orders, data transfers smoothly, without disruptions or errors. It significantly boosted efficiency across all the company’s business processes»
Role Model Customization
Complete control and business scalability without expanding staff or compromising on operational quality
Task Segmentation into Roles with Different Access Levels
Automation went beyond just technical integrations. The company also implemented a flexible and scalable role model for employees, with varying levels of access, qualifications, and responsibilities. Some roles could only perform basic tasks (like scanning items in the warehouse), others handled creating shipments, and yet others managed inventory and pricing. This made it clear who was responsible for what and significantly reduced the risk of mistakes; this not only optimized staffing but improved overall process transparency and security.
«Thanks to role customization, we’ve eliminated the need to use the MP Personal Account altogether. Now we know exactly who’s responsible for each step, and we can monitor any staff actions»
Introducing a Night Shift with Less Qualified Staff
A special solution made possible by the flexible role model was the introduction of a night shift. The company hired employees with minimal qualifications for night work in the warehouse. Thanks to well-defined roles and an automated system, these employees could perform their tasks accurately despite limited knowledge of internal processes.
«We significantly reduced payroll costs by creating a night shift where employees carry out only simple tasks that are nearly foolproof. This not only saved the company money but also noticeably cut down on errors»
Through automation and the role model, the company dramatically reduced mispicks, mistakes in assembling orders, and marketplace penalties. Every step in order processing was now transparent and traceable. All staff actions were logged automatically, allowing any error source to be quickly identified and addressed, while also giving management real-time insight into staff efficiency.
Order Processing Customization
Solving key operational challenges in quality control and order accuracy
As the seller operates on the scale of a major retailer, the standard approach to order handling wasn’t sufficient. They decided to develop and implement a custom workflow that suited the specifics of an Enterprise business and helped them avoid common operational errors.
Two-Level Order Check
One critical innovation was a two-level check for each order. Every order underwent two mandatory scanning steps:
- Initial scanning during warehouse picking to confirm item selection and order accuracy.
- Repeat scanning just before shipping the batch of orders to Wildberries, identifying any errors or missing items before the package went out to the customer.
This approach significantly improved order processing quality and accuracy:
«With two-level verification, we minimized the risk of incorrect items or lost goods. Even if a mistake slipped through during initial picking, the second scan would catch it before the order reached the customer»
Built-In Protection Against Mispicks and Lost Items
Beyond two-level checks, the order orchestrator also included automated safeguards against the most common warehouse errors, such as mispicks (sending the wrong item) and lost packages. These safeguards alerted staff to potential issues and automatically triggered a recheck process if needed.
Automating Stock and Margin Management
The next important step was automating inventory and pricing management. The company built a system that tracked price and stock changes in real time, calculated margins, and quickly decided whether products should remain listed on Wildberries.
Real-Time Price Management System Integration
They integrated their Master Price System with the order orchestrator. Any price changes were instantly shared with the order management system, preventing sales based on outdated or incorrect prices.
«Prices can change daily, or even more frequently. Thanks to integrating Master Price System, we always ensure that the prices shown on Wildberries are accurate, consistent with our internal strategy, and aligned with our planned margins»
Auto Margin Calculation and Reactive Stock Nullification
Another key feature was automatic margin checks for every item. In real time, the system would calculate each product’s profitability based on marketplace fees and current cost. If an item’s margin fell into negative or unacceptably low territory, the system would immediately zero out its stock, removing the product from sale. This same mechanism supported exceptions for clearance items.
«If an auto-discount kicked in and made a product unprofitable, our system caught that right away and removed the item from sale. This let us protect the company’s profit and avoid selling at a loss»
Using «Gaps» for Safe Inventory Management
To mitigate «out of stock» risks, the company introduced special «gap» mechanisms. This means the marketplace typically sees a lowered stock count. Even with a sharp spike in demand, these gaps give the seller time to update inventory data and make quick decisions.
«Gaps became our shield against unexpected demand surges and overlapping sales from other channels. Thanks to this mechanism, we almost entirely eliminated order cancellations due to lack of stock»
Developing a Custom PIM for Content
One of the most critical tasks for the seller was managing a massive product catalog and adapting it to the marketplace’s requirements. To do this, they built a specialized PIM (Product Information Management) system designed specifically for marketplace integrations.
Centralized Product Listing Management, Aligned with Wildberries Requirements
The PIM system was built to simplify preparing and publishing content on Wildberries. It included fields and attributes fully matching the marketplace’s specifications. This let the system automatically format the company’s internal data (color, size, product description, etc.) to meet Wildberries standards, without manual edits for every product card.
«Our standard PIM — used for offline retail and our own website — wasn’t cut out for Wildberries. We had to develop a specialized system with all the unique attributes and fields the marketplace demands. Thanks to that, we reduced the time to launch new products and nearly eliminated manual errors»
Minimizing Manual Edits and Errors When Listing Products
Previously, listing each product required extensive manual revisions and checks by the content department. After launching the specialized PIM, most of these tasks were automated. The system converted the company’s internal descriptions, colors, and sizes into a format accepted by Wildberries.
Automating Accounting Reports
A vital part of the comprehensive automation effort was improving the accounting process. With so many transactions coming through Wildberries, the company needed a mechanism to automatically generate accounting reports.
Nightly Report Generation with Minimal System Load
The company set up automated reports to run overnight, when infrastructure usage was low. These reports compiled all sales, returns, and financial data, greatly simplifying the accounting department’s workload and eliminating the need to manually export data from the Wildberries dashboard.
«Previously, accountants spent a huge amount of time manually exporting reports. Now, when they come in each morning, they have all the data prepared, aligned with our internal standards and formats. This dramatically speeds up and improves the accuracy of accounting»
Plan for Full Integration with Accounting Systems
Next, Desport intends to integrate the reporting system completely with a major accounting solution, which would eliminate even the slightest manual intervention.
«Our next objective is a fully automated connection to our accounting systems, so that accountants can forget about manual reconciliations or exports. This will make their jobs easier and further boost the speed and accuracy of our financial reporting»
Technical Integration Details
To ensure scalable and reliable business processes with Wildberries, Desport built a high-performance, stable architecture using modern technologies and proven solutions. This section covers the key technical details and the specific implementation features.
Solution Architecture
The integration architecture was centered on the order orchestrator.
In essence:
- WB API transmitted data about new orders, returns, and shipments to the orchestrator.
- The orchestrator continually polled all systems for order status updates and routed this information using Apache Kafka to the ERP system.
- The ERP system, in turn, exchanged order and stock data with the warehouse management system.
- The Master Price System automatically updated prices and calculated margins.
- The BI system received analytics and data for reporting and forecasting.
«We designed the integration so that every component operates in real time. All information from Wildberries reaches the warehouse and ERP within seconds, allowing us to react immediately to any changes in orders or stock levels»
Technologies Used
They chose the following technologies to deliver a high-performance, reliable solution:
- Apache Kafka — the main tool for high-load data exchange, replacing traditional REST APIs and significantly speeding up data transfer between the orchestrator and ERP.
- Memcached — used to match the seller’s internal order IDs with Wildberries order IDs. This sped up processing for thousands of orders and reduced database load.
Problems Encountered and How They Were Solved:
Initially, the company struggled with SAP ERP REST API overload from high traffic and frequent requests. As a result, they decided to switch to Kafka, which solved the capacity problem.
They also initially faced challenges linking order numbers quickly, which Memcached resolved. This improved order handling performance and lowered the risk of mistakes.
«Switching to Kafka and Memcached removed the main technical limitations of our infrastructure. We ended up with a system that can handle extreme loads without data loss or delays»
Implementation Highlights
Ilya shared some specific solutions that helped the system maintain its quality.
Algorithmic Checks
The orchestrator included specialized real-time algorithms that automatically verify correct order assembly, cutting down on the likelihood of mispicks or lost products in the warehouse. If any issues were detected, the algorithms automatically initiated additional checks.
Working with Chestny Znak (Marking System)
A special technical challenge was managing labeled products, which require mandatory registration in the «Chestny Znak» marking system. To handle this, the company implemented a dedicated gateway that communicates with «Chestny Znak» Every labeled item is automatically registered and validated at each movement and sale, ensuring compliance with legislative requirements.
«Product labeling is very serious, so we built a separate integration with the ‘Chestny Znak’ system. It works at the WMS picking stage: each label code is captured as an attribute of a specific SKU, then passed on to all key systems. We can guarantee that the label’s status is always up to date and meets legal requirements. Yes, the rollout was tough, but the results fully justify the effort»
Results and Achievements
Increased operational efficiency, protected margins, and expanded geographic reach
Through comprehensive API integration and automation of key business processes, Decathlon Russia—now Desport—achieved impressive outcomes on Wildberries. By implementing custom solutions, the company substantially grew sales volume, improved efficiency, maintained tight margin control, and expanded to new regions.
Financial Performance
The company started with Wildberries essentially from scratch. Within the first three months, they posted remarkable achievements:
- Revenue jumped from zero to hundreds of millions of rubles in just 3 months.
- Over the following calendar year, the company saw its revenue grow by nearly 6 times.
- Average daily sales on Wildberries matched those of Decathlon’s largest offline flagship stores.
- For a certain period, sales through Wildberries reached about 5% of Decathlon’s overall business in Russia.
«When we were just starting, we couldn’t have imagined these numbers. Thanks to API integration and process automation, our Wildberries sales reached the level of our biggest stores, and during sale events, we even surpassed them multiple times»
Operational Efficiency
Automation had a strong impact on day-to-day efficiency:
- The staff needed for order processing was cut by roughly 40% compared to the pre-MVP phase, owing to advanced process automation and role-based workforce organization.
- Nearly all human errors in order processing were eradicated, dramatically reducing penalties, mispicks, and lost products.
«Previously, we had about 80 people manually processing orders. After automating and adopting a role-based model, we almost halved the team and moved a lot of tasks to a night shift with less qualified and cheaper labor, without losing quality»
Margin and Pricing Control
Automated stock and pricing management gave the company complete control over its financial performance on Wildberries:
- Substantially reduced losses from auto-discounts that used to eat into margins.
- Implemented a system to manage stocks and prices, automatically pulling products if margins dropped below a set threshold.
«We solved the auto-discount problem by building a system that immediately responds to margin changes. Our products always remain profitable, and we no longer lose money on unwanted discounts»
Geographic Reach
By automating with Wildberries, the company significantly expanded its sales territory, reaching regions previously inaccessible through traditional channels:
- 80% of Wildberries orders came from remote areas of Russia, where Decathlon has no brick-and-mortar presence.
- The marketplace gave the company access to a customer base that would have been difficult or too expensive to reach alone, opening up entirely new markets and growth opportunities.
«We studied the data closely and found that 80% of our Wildberries buyers live over 200 km away from our nearest stores. That proved we weren’t cannibalizing ourselves, but rather tapping completely new customers»
Insights and Conclusions
After extensively integrating and automating its Wildberries operations, Desport gained valuable experience that can guide other major retailers looking to scale effectively on Wildberries or other marketplaces. Key lessons learned include:
A Role-Based Model Is Crucial for Scaling
Clear segmentation of staff responsibilities and access levels was vital to Desport’s success in scaling. A properly configured role model reduces human errors, lowers operational costs, and brings transparency to every business process.
Automation Is the Best Path for Enterprise
Desport’s experience shows that manual methods simply aren’t viable with high order volumes and large product ranges. Only true automation ensures the level of quality and stability required by Enterprise operations.
Understanding Marketplace Requirements Offers a Competitive Edge
Marketplaces enforce unique rules for product content, order processing, and returns. The sooner a company accounts for these rules in its systems and processes, the faster it can adapt and the less it will spend correcting mistakes.
Accounting and Returns Are the Next Improvement Areas
Desport also highlights the ongoing need for better integration with accounting and more automation for handling returns. A robust automated data flow with accurate order identification significantly simplifies and speeds up operations.
«Automation is an ongoing process. We won’t stop here and will keep enhancing our infrastructure to further improve quality and profitability»
Advice for Sellers
The company’s experience vividly demonstrates that with the right approach and well-designed API integrations, large enterprises can effectively and profitably scale their marketplace sales without raising operational costs or risks.
For those still on the fence about whether to automate their processes and integrate via APIs, consider the key advantages Desport gained from these solutions:
- Significant revenue growth (nearly 6x in one year).
- Noticeable cut in operating costs (about 40% fewer staff).
- Higher accuracy and fewer losses due to discounts or mispicks.
- Wider geographic coverage and entry into new markets.
«Don’t wait until manual processes become a bottleneck. API integration has already proven vital and profitable for businesses of any scale. Start automating today to secure a competitive edge tomorrow»
— Ilya Botvinnikov, Desport
Appendix
The most important WB API functions mentioned in the article:
