Study on the evaluation of multimedia advertising performance

With the development of multimedia information technology, all walks of life use multimedia technology to publish business information and advertising information to enhance users' awareness and satisfaction with the industry. Among the applications of using multimedia to enhance management, keyword advertising is undoubtedly one of the successful models. Keyword advertising, a typical kind of multimedia advertising, is an important way for advertiser to match customer needs precisely. Evaluating the effectiveness of this novel advertising approach comes to be the focus of attentions in this field. In addition, user experiences in advertisers' page are hardly taken into account in existing researches. Using a unique dataset of keywords on a commercial search engine provider in China, we model click, cost per click and conversion simultaneously. We introduce factors of three categories: bidding factor, keyword semantics factors and user cognition factors. Page view and bounce rate, which have not been reported in existing researches as far as we know, are employed to discover user cognition on website. We find that cost per click has a significant relationship with clicks and conversions. Brand information can increase both clicks and conversions while decrease cost per click. Although long tail keyword advertisement may bring website more traffic, it does not increase conversions. Specifically, we find that page view is positive relate to clicks and conversions while bounce rate is negative to conversions. Our research results can be easily applied in practice by advertiser to improve multimedia advertising performance.