BTP Score – Our pet project
One of the wonderful feature of the Amazon product page is showing what people bough eventually after viewing that particular item. As a consumer, we have a limited idea on which exact model of a series of products is really doing well and which isnt. The knowledge we get is derived from reading reviews and feedback from friends and family.
Me and Nithin (my CTO) were thinking on how better can we provide a metric that goes beyond the user ratings/reviews in helping a visitor gauge the product better. We believe that there is a lot more information that a little number crunching will reveal.
For example, the average daily revenue from a product is a good indicator of its contribution index. Ideally, if that particular item is selling more, it should be made more visible to the customer by getting featured on better real estate. The number of searches a particular product receives via the BTP search box is an indicator of its popularity index. In all we decided to use the following 7 parameters with the right weightage thrown in to arrive at a score on 100.
(s) number of times it is searched
(q) quantity sold
(p) price of the product
(d) days since live
(v) visits
(uv) unique visitors
(t) time on page
Of course, a lot of tweaks had to be done to the formula for it apply to the entire spectrum of products. For ex, we cant take the revenue factor on face value for a Rs 30,000 product Vs Rs 300 product. Finally, we seem to have understood and arrived at a formula that does true justice. We would be unveiling it in a week, and cant wait to understand if it proves right – should ideally show in the trend of sales vs scores ! Shall keep you posted !