Data-driven influencer marketing strategy analysis and prediction based on social media and Google Analytics data

Abstract

Due to various uncertainties on social media, data-driven strategy has become a necessity for influencer marketing. Typically, a promotional post by an influencer aims to direct the viewers to buy a product from a brand's website. The objective of this paper is to analyse the factors that contribute to the popularity of promotional posts in terms of likes and website visits count. This research utilised Facebook (FB), Instagram (IG) and Google Analytics (GA) data collected from the ambassadors (or influencers) of MeCan App, a Malaysian e-commerce company. The factors that contribute to popularity have been successfully identified, such as the optimal posting time, hashtags, image type, interval and ratio of posts. For example, they should post based on the ratio of one regular post for 5.4 promotional posts for the best exposure. Additionally, regression methods were implemented to predict website visit count, with an accuracy of 69.9 per cent using Random Forest Regressor.