The company runs a large online auction platform and shopping website that connects buyers and sellers in over 190 countries. Their website has millions of sellers with 1.9 billion global listings and over 132 million active buyers.
THE CHALLENGE
The sellers on the Company’s platform were a diverse lot and varied in terms of value of products (Gross Merchandise Value or GMV) sold, quantity of products sold, type(s) of products sold, geographies, etc.
To identify the high-risk sellers, our customer classified them into broad groups. Each group had its own GMV and quantity limits. The company had a provision to allow the top performers in each group to shift to the next group with higher GMV and quantity limits.
The sellers which indulged in high-risk and / or unfair practices were given warnings, with repeated infractions leading to their accounts being suspended and then banned. However, some of these sellers would create new accounts and continue to sell the same products on the company’s platform.
Overall, the company needed an analytics partner to develop a comprehensive risk management solution which defined the different seller categories and their linked GMV and quantity limits, included a risk scoring system based on various parameters, identified and flagged linkages between duplicate accounts created by the poorly performing sellers and so on. This solution would help the company get a comprehensive view of the performance of their sellers and quantify the risks associated with their bad sellers.
THE SOLUTION
The team of business analysts from Prescience Decision Solutions assessed the existing seller categorization process and the corresponding monthly GMV and quantity limits for each group. After selecting a sample set of sellers, our team analysed a wide range of details such as the Know Your Customer (KYC) status, adherence to the company’s seller registration processes, overall performance of the seller, customer feedback on the ordered goods, delivery process and seller interactions, and so on. Based on all these factors, the Prescience team designed a new complex risk scoring system, with a higher score indicating a better performing seller.
The existing system of categorizing sellers was replaced by a newly created system with 8 separate groups. Each group had a fixed GMV and quantity limit. Depending on the risk score of each seller, they were automatically assigned to their corresponding risk group.
With these clearly demarcated risk groups, the company was able to consistently track seller performance over a period of time. The best sellers in each category were able to apply for a move up to the next level with greater GMV and quantity limits, while the sellers who continued to underperform were tagged to a higher risk category group with lower GMV and quantity limits.
The solution was capable of identifying the new duplicate accounts created by high-risk sellers who had negative reviews and were flagged for poor performance. Using A/B testing, the company could also monitor seller behaviour in each risk category including which sellers were listing out their entire inventory and which other sellers were deliberately withholding products from their marketplace.
THE IMPACT
With the updated risk management solution, the company was able to create specific new risk categories which were in line with their business requirements. The new seller risk scoring system allowed the company to accurately gauge the probability of risk with each seller and assign them to the updated risk categories.
The long-term behaviour and performance of all the sellers was easily tracked with this risk management system, with high performing sellers being rewarded by moving them into lower risk categories with higher GMV and quantity limits.
