Get a free consultation with a marketer marketolog@adbest.ru
Even at the descriptive statistics stage, analytical marketing research often reveals very interesting and significant results, which can be further used as initial hypotheses for more complex methods. For example, it may turn out that men and women behave differently on a website, are interested in different information, or that an above-average level of solvency is typical of website visitors. Or, for example, that a particular page on the website has an abnormally high level of visitors, despite it not being the company's best-selling product. The chi-square test allows us to understand whether the relationship between categorical variables is significant.
Go to catalogRegression analysis allows us to delve much deeper into the cause-and-effect relationships of phenomena. For example, in contextual advertising, when we encounter the problem that ads simply aren't showing when advertising is enabled, we conduct a regression analysis to determine which significant factors influence the number of impressions and obtain a list of pages that will actually show better. If we have a fixed budget, we can determine which parameters will maximize the return on advertising. If we have only categorical variables, we can still answer the question of whether a customer will perform the desired action. We simply use other marketing analysis methods that don't require all variables to be quantitative, such as discriminant analysis or decision tree construction. Mathematical statistics methods allow us to select those parameters that, all other things being equal, will yield the greatest return, and to identify those niches and products that are of greatest interest to consumers and worthy of dedicated promotion and branding..
Modeling contextual advertising parameters using statistical methodsOur projects
