Učinkovitost spletnega oglaševanja pri uporabi vedenjskega ciljanja uporabnikov

STRUNA, SIMON (author), Bosnić, Zoran (mentor)

V našem diplomskem delu smo želeli na podlagi predpostavk vedenjskega ciljanja in sorodnih raziskav preveriti, ali vedenjsko ciljanje res lahko izboljša učinkovitost spletnega oglaševanja. Za namene raziskave smo pridobili anonimizirane podatke iz slovenske oglaševalske mreže. Vsakega uporabnika smo na podlagi obiskanih strani in kliknjenih oglasov predstavili v obliki uporabniškega vektorja in vzporedno izvajali dva neodvisna eksperimenta. Izvedli smo postopek gručenja, nato pa za posamezne oglase poiskali najprimernejše gruče ter simulirali prikazovanje oglasov zgolj tem gručam. Primerjali smo dvig učinkovitosti oglasov pred in po gručenju ter hkrati preverjali še karakteristike podobnosti. Pokazali smo da lahko dvignemo učinkovitost vedenjsko ciljanih spletnih oglasov za kar746 %. Na koncu smo rezultate in ugotovitve primerjali še z izsledki iz sorodne literature in v zaključku podali nekaj predlogov za izboljšavo naše metode in ideje za nadaljnje delo.

• Language: Slovenian
• Keywords: oglaševanje, vedenjsko ciljanje, oglasi, gručenje
• Work type: Undergraduate thesis
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2014
• PID: 20.500.12556/RUL-30042
• Publication date in RUL: 14.11.2014

Secondary language

• Language: English
• Title: Efficiency of web advertising with behavioral targeting of users

Abstract

In our thesis we wanted to examine, based on assumptions of behavioral targeting and related research, whether behavioral targeting can really improve the effectiveness of online advertising.We conducted our experiment on anonymised information that we have obtained from a Slovenian advertising network. Based on previously visited web pages and clicked ads, we represented each user profile as a vector of interests and carried out two independent experiments. We implemented the clustering method, simulated displaying ads to different clusters and compared the increase in the effectiveness of advertisements before and after clustering. We proved in our experiment that using behavioral targeting and short-term data for capturing user's interests, efectiveness of online advertisements can be raised as high as 746 %. Finally, we compared our results with the results of related literature and provided some suggestions to improve our methods and ideas for future work.

• Keywords: advertising, behavioural targeting, ads, clustering