Evaluation of effectiveness of an advertisement campaign by use of Twitter

Abstract

The paper addresses the topic of marketing campaign evaluation by means of statistical analysis of publicly available data from popular micro-blog service Twitter. The research is aimed at finding out the parameter or combination of parameters e.g. average frequency, peek frequency, total number of posts, number of positive/negative posts etc. of tweets mentioning the product/company of interest providing best correlation with sales. In particular current paper addresses movie industry by correlating first weekend box office revenue of all movies premiered in the period from September 2012 till end of March 2013 in USA with the amount of tweets mentioning the movie titles during Monday to Sunday in the week of first night of a movie. Moreover analysis is given for “positive tweets” and “negative tweets” search options of Twitter and their effect on the prediction. Additionally the paper outlines major general issues associated with data collection and analysis using Twitter.