The corporation runs a large online auction platform and shopping website. It has over 170 million active buyers globally and one billion active listings. It has a significant amount of transaction history and processes more than 100 petabytes of data daily.
The company runs approximately 150 promotions a year targeting more than one million active sellers, and thereby contributing significantly to the Gross Merchandise Volume (GMV). However, the existing promotions were not proving to be as effective as desired by both the company and the sellers. There was an urgent need to customize promotions by optimizing the targets and the offers in order to improve net revenue and ROI.
In addition to these problems, the number of sellers active on the e-mail marketing channel was remarkably low. Hence, the number of opt-in sellers was minimal in spite of an attractive offer.
There was a pressing need to leverage the “active” seller segment to improve net revenue.
Our business analysts classified the sellers into different segments based on their listings, GMV contribution, average selling price of items, quantity of items sold and their behaviour before and during the promo-periods.
Each seller was scored based on “propensity to list with a promo minus propensity to list without a promo” using various modelling techniques. These scores helped us in identifying the seller segments that would anyway list on the website and also the sellers who would list more often with a promotion in place.
Our solution was also able to identify the sellers who list only with a promotion and don’t list without a promotion, based on their opt-in behaviour in the preceding six months.
The client was able to reduce costs and and improve their ROI significantly, by excluding sellers who would list on the website irrespective of receiving a promotion. The client was able to see the difference in how the sellers who took advantage of promotions also helped in their cost reductions. The solution also improved the behaviour of sellers during a non-promotion period.