Optimization of Paid Search Traffic Effectiveness and Users’ Engagement Within Websites

Abstract

Optimized paid search advertising campaigns composed of multiple data analytics insights and prior experiences of search engine marketing performances. However, when marketers compete in the battle of paid search ads’ rankings, complexity in optimization is increased. The higher the search ads’ ranking position, the greater the chance that users of search engines will click the ads. Despite the existing knowledge of the factors that contribute to the higher ranking position in search ads, such as proper relevancy among users’ search terms and text ads, or landing pages content, little is known about search engine users’ behavior after ads clicking. Low interaction or immediate abandonments from the landing pages potentially lead to a waste of budget spent on each paid advertising campaign. In this regard, marketers should pay much more attention to the interaction of paid traffic visitors after clicking on search ads, and not only to search engine rankings and user impression share rates. In this paper, the authors develop a computational data-driven methodology with a purpose to estimate and predict paid traffic visitors’ engagement in seven courseware websites after clicking on the search ads. The higher the engagement with the landing page, the higher will be the probability for conversions. At the first stage, web behavioral analytics are retrieved for 120 consecutive days in certain web metrics. These are the volume of paid traffic visitors, the average pages per session, the average session duration, and the bounce rate. Statistical analysis of the extracted web behavioral datasets takes place for understanding the cohesion, validity, and intercorrelations between the web metrics. KMO and Bartlett’s test of sphericity and Pearson coefficient of correlation are adopted. One step further, agent-based modeling and simulation is adopted as a methodology for abstracting and calibrating paid traffic visitors’ behavior inside the examined websites. Poisson distributions are implemented for predicting the potential engagement of paid traffic visitors in specific date ranges. Through this, the paper highlights its practical contribution to marketers with the purpose to develop search engine marketing campaigns composed of search ads relevant to the users and sufficient content engagement after ads clicking.KeywordsDigital marketingSearch engine marketingSearch advertising optimizationWeb analyticsBehavioral analyticsData-driven modelingAgent-based models in marketingModeling and simulation in marketing