Sellers across the country list products of various categories like apparels, footwear, fashion accessories, etc. via a self serve UI, provided by one of the leading e-commerce portals in India. Over 20% of catalog attributes provided by these sellers are blatantly incorrect resulting in high customer returns and dissatisfaction. Hence, a stringent QC process for incoming catalog is imperative before it goes live on the e-commerce portal.
SCALE: Checking and correction of 5000 products per day
Expected TAT: 24 hours
Client integrated with Playment’s APIs to facilitate automated data transfer on realtime basis.
Playment built a completely automated workflow consisting of two major steps - Verification & Correction.
Verification - 20 different types of tasks, as part of workflow, were created to check 5-6 attributes (Category, Brand, Colour, Print type etc.) for products spanning across 60+ categories. The incorrect values were flagged.
Correction - Flagged incorrect values were rectified by via expert Players on the platform.
Tasks were routed to only expert Players in cataloging exercises, selected using Playment’s self-learning player selections logic (PSL) which evolves every time a Player submits a task.
Depending upon level of subjectivity a particular task was routed to 3 - 7 expert Players on the platform until the most accurate answer is sourced.
Final output was provided back to the client in the desired format through an API.
We plan to increase the scale from 5,000 products/day to 1,00,000 products/day as we constantly improve the overall accuracy and reduce TATs.
Request Demo to know more.