izpis_h1_title_alt

Optimizacija konverzij v programatičnem oglaševanju
ID ANDROJNA, LUKA (Author), ID Šubelj, Lovro (Mentor) More about this mentor... This link opens in a new window, ID Tori, Andraž (Comentor)

.pdfPDF - Presentation file, Download (1,10 MB)
MD5: 0EF4AAF4408F9058A1B6FB201ED6106D

Abstract
Založniki na spletu na svojih spletnih straneh služijo z oglaševanjem. Oglasi so namenjeni bralcem vsebin založnikovih strani, za njih pa plačujejo oglaševalci. Ti ponavadi prikazujejo oglase, da bi izboljšali ugled znamke ali pa pridobili uporabnike, ki opravijo določene akcije. Oglaševalci si želijo, da preko klika na oglas k njim pride obiskovalec, ki jim bo nekaj doprinesel, na primer ustvaril nov uporabniški račun, se prijavil na njihovo glasilo, opravil nakup itd. Takim akcijam pravimo konverzije. Cilj programatičnega oglaševanja je optimizirati nakupovanje oglasnega prostora tako, da se zgodijo kliki na oglas in konverzije. V tej nalogi si bomo ogledali različne pristope, kako znižati ceno za akvizicijo in povišati razmerje med kliki in konverzijami. Problema se lotimo z uporabo seznama dovoljenih kombinacij oglasne borze, spletne strani in identifikatorja oglasnega mesta ter seznamom prepovedanih uporabniških agentov. Z A/B testom, ki je potekal v produkcijskem okolju in vplival na kupovanje v realnem času podjetja Zemanta d.o.o., smo potrdili, da s kupovanjem prometa na boljših oglasnih mestih pridobimo boljše razmerje med kliki in konverzijami ter nižjo ceno za posamezno konverzijo. Opazimo tudi, da je niˇzja tudi cena klika, čeprav je cena prikazov višja.

Language:Slovenian
Keywords:oglaševanje, podatkovna znanost, podatkovna analiza
Work type:Bachelor thesis/paper
Organization:FRI - Faculty of Computer and Information Science
Year:2019
PID:20.500.12556/RUL-106524 This link opens in a new window
Publication date in RUL:01.03.2019
Views:951
Downloads:223
Metadata:XML DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:Optimization of conversions in programatic advertising
Abstract:
Publishers on the web monetize their businesses by running advertisment on their websites. Ads are targeted at their readers and paid for by advertisers. Advertisers generally show ads to either improve their brand perception or to get users to make some specific action. They want to get engaging users onto their product’s landing page and take actions that generate value for the advertiser, for example registering a new account, subscribing to a newsletter, making a purchase etc. Those actions are called conversions. Optimizing buying of advertising to induce clicks and conversions is the business of programatic advertising. We will look at different approaches for lowering the CPAs and raising conversion rates. We approach the problem with the use of a whitelist containing combinations of exchanges, publisher websites and tag ids, and a blacklist of useragents. We conducted an A/B test, which was running in an online production environment and affected bidding in real time for Zemanta d.o.o., that showed buying traffic on premium ad space gives us better conversion rates and lower prices per acquisition. We also noticed that prices per click dropped even though the average price per impression went up.

Keywords:advertising, data science, data analysis

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back