Keyword bidding optimization strategy on e-commere platforms

Abstract

The primary objective of this study is to understand the process of keyword bidding in ecommerce platforms and develop a strategy for the sellers, so that they can optimize their keyword bidding strategy. The formulated strategy is aimed at increasing the visibility and ranking of the seller’s products in the search results of e-commerce platforms. This will subsequently increase the chances of more impressions, click through rate and conversions. We have done extensive research on current practices used for keyword bidding, i.e., automatic, and manual bidding and the various scenarios of obtaining monetary benefits for the sellers. The various terminologies used in keyword bidding are explained and the different performance metrics used to measure the effectiveness of the bid keywords are provided in this report. We have analysed the need for considering different strategies for different products of the sellers and the disadvantage of having one blanket strategy for all products and at all times of a given day. The report provides idea about different types of matches between the search term of the users and the keywords bid by the sellers. We have analysed the importance of the match types when it comes generating impressions and the relevant match types suited for different scenarios like product popularity, brand awareness, degree of competition for a particular product, etc. Next section of our report covers the steps in setting up a campaign and the constraints and resources needed to be considered in each step of the campaign. Based on the research done and inferences made, we have proposed our idea of how sellers should go about their bidding strategy during high and low competition scenarios. This kind of analysis is required for start-ups as they need to be very careful with their keyword bidding approach. The strategy required for start-uos and established companies have been compared and analysed and our insights have been reported. We have incorporated analytical approach in our study in the form of panel data analysis. Regression model has been extensively used to predict the number of impressions or conversions or clicks. This can be done given the availability of information of how each and every keyword has been performing at different times in a day and the amount spent on bidding that keyword.