80% of the sales conversion is attributed to the suggested search results. Search is the most powerful feature to direct a customer to the most relevant product he/she is searching for, increasing the chances of conversion drastically. An Indian e-commerce giant wanted to fine tune its search and recommendation algorithms to amplify sales conversion. This is a perpetual activity and requires continuous monitoring of search results for the smallest of the change in algorithm structure.
SCALE: 5000 queries per day
Expected TAT: 24 hours
- The leading e-commerce giant provided the search queries and the top 10 results for each query.
- Each set of 10 results were to be classified as Good, Average, Poor depending upon the relevance of the fetched products by the query, appearance of out of stock/obsolete products and 5 more such parameters.
- Detailed instructions were sketched out to train Players for judgment.
- Playment trained its Players with the help of Tutorials and Qualifier Tasks as per the instructions for all the judging parameters.
- After selecting the expert Players from the qualifiers, they were further classified based on their personas. More Tasks were provided to the Players who shop online more frequently.
- The opinion of minimum 7 Players were recorded for each set of search results.
- An aggregated score on the scales of 3 was provided to the Partner for the set of 5000 queries.
- Tracking this score for a new set of queries on a weekly basis helped the Partner to gauge the performance of the deployed algorithms and take preventive steps accordingly.