The online advertising market is continuously evolving, with an estimated forecast of companies investing nearly 600 billion dollars in online advertising campaigns by 2023. The success of this channel lies in marketers' ability to target specific users for directing advertising messages, encouraging them to make online purchases or perform other actions. To guide users toward sales, marketers create advertising campaigns across various channels and platforms, each aimed at a precise target with a specific message and daily budget. Marketers typically aim to allocate advertising investments across these campaigns to maximize overall profit. However, it is often challenging for companies to achieve high sales volumes without compromising return on investment (ROI) targeting a limited audience. In this thesis, we introduce TargExp, an algorithm designed to broaden the target audience of advertising campaigns to achieve greater volumes and profits in online advertising. The proposed approach comprises three phases: audience identification, where analytical APIs and Scraper tools are employed to discern user characteristics; performance estimation, focusing on data-driven insights to estimate campaign effectiveness; and budget allocation for profit maximization, ensuring maximum profitability by investing in the most promising campaigns. The proposed approach has been implemented in a SaaS platform and released in a real environment. Initial tests of the algorithm show extremely promising results, demonstrating in an experiment involving over 15 real campaigns an increase in profit of over 48.
Il mercato dell'advertising online è in costante evoluzione, con una spesa stimata di oltre 600 miliardi di dollari in campagne pubblicitarie online entro il 2023. Il successo di questo canale risiede nella possibilità per i marketer di mirare a utenti molto specifici per indirizzare messaggi pubblicitari personalizzati, incoraggiandoli a effettuare acquisti online o più genericamente generare conversioni (es., iscrizioni a newsletter, compilazione form, richiesta di un preventivo ecc.). Per guidare gli utenti verso gli acquisti, i marketer creano campagne pubblicitarie su vari canali e piattaforme, ciascuna mirata a un target preciso con un messaggio specifico e un budget giornaliero. Il goal dei marketer è quello di massimizzare il profitto complessivo di queste campagne distribuendo in modo efficace gli investimenti pubblicitari. Tuttavia, mirando a un pubblico limitato, spesso è difficile per i team di marketing raggiungere volumi di vendita elevati senza compromettere vincoli relativi al ritorno sull'investimento (ROI). In questa tesi, presentiamo TargExp, un algoritmo progettato per ampliare il pubblico target delle campagne pubblicitarie per ottenere maggiori volumi e profitti nella pubblicità online. L'approccio proposto comprende tre fasi: identificazione del pubblico, dove vengono impiegati strumenti di analytics, API e scraper per identificare cluster di utenti promettenti; stima delle performance, al fine di stimare le performance attese delle performance di campagne pubblicitarie indirizzate a tali utenti; e allocazione del budget per la massimizzazione del profitto, garantendo massima redditività investendo nelle campagne più promettenti. L'approccio proposto è stato implementato in una piattaforma SaaS e rilasciato in un ambiente reale. I test iniziali dell'algoritmo mostrano risultati estremamente promettenti, dimostrando in un esperimento con oltre 15 campagne reali un aumento del profitto di oltre il 48%.
TargExp: an Algorithm for Audience Expansion and Profit Maximization for Online Advertising
EL KHOURY, JANA
2022/2023
Abstract
The online advertising market is continuously evolving, with an estimated forecast of companies investing nearly 600 billion dollars in online advertising campaigns by 2023. The success of this channel lies in marketers' ability to target specific users for directing advertising messages, encouraging them to make online purchases or perform other actions. To guide users toward sales, marketers create advertising campaigns across various channels and platforms, each aimed at a precise target with a specific message and daily budget. Marketers typically aim to allocate advertising investments across these campaigns to maximize overall profit. However, it is often challenging for companies to achieve high sales volumes without compromising return on investment (ROI) targeting a limited audience. In this thesis, we introduce TargExp, an algorithm designed to broaden the target audience of advertising campaigns to achieve greater volumes and profits in online advertising. The proposed approach comprises three phases: audience identification, where analytical APIs and Scraper tools are employed to discern user characteristics; performance estimation, focusing on data-driven insights to estimate campaign effectiveness; and budget allocation for profit maximization, ensuring maximum profitability by investing in the most promising campaigns. The proposed approach has been implemented in a SaaS platform and released in a real environment. Initial tests of the algorithm show extremely promising results, demonstrating in an experiment involving over 15 real campaigns an increase in profit of over 48.| File | Dimensione | Formato | |
|---|---|---|---|
|
2023_12_ElKhoury_Executive_Summary_02.pdf
non accessibile
Dimensione
514.86 kB
Formato
Adobe PDF
|
514.86 kB | Adobe PDF | Visualizza/Apri |
|
2023_12_ElKhoury_Tesi_01.pdf
non accessibile
Dimensione
1.25 MB
Formato
Adobe PDF
|
1.25 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/214176