PROMOS
PRediction and Optimization of MObile Subscription marketing campaigns
Project Designation
PROMOS - PRediction and Optimization of MObile Subscription marketing campaigns
Reference
NORTE-01-0247-FEDER-017497
Main Objective
Strengthen research, technological development and innovation
Funding
Co-Promotion R&D Project, Aviso 33/SI/2015 - I&DT Empresarial, funded by Portugal 2020/Adi-Innovation Agency
Promoter Entity
OLAMOBILE PORTUGAL, SOCIEDADE UNIPESSOAL, LDA
Co-promoter Entity
University of Minho
Investment
730,163.82 EUR
Elegible
721,888.81 EUR
Non Repayable Subsidy
534,610.22 EUR
Duration
December 2016 - November 2019
Project Description
This project involves the area of Mobile Performance Marketing (e.g. smartphones, tablets) and advertising of mobile products with a subscription business model. It is an advertising growing area due to the evolution of the Internet and mobile market and it corresponds to a marketing business where the advertiser only pays when there are acquired customers and measurable results (e.g. Cost Per Acquisition - CPA). In particular, an automatic algorithm will be researched and developed for the PRediction and Optimization of MObile Subscription marketing campaigns (PROMOS). The aim is to predict what is the best mobile product to be shown to the end user, while optimizing the interests of those involved in the business: OLAmobile company (co-promoter, whose main activity is related with Mobile Performance Marketing), advertisers, webmasters and creators of mobile content and also users.
The algorithm will be studied (researched, developed and tested) in a laboratorial environment (Technology Readiness Level - TRL 4) by the co-promoter University of Minho (UMinho), being based on a self-learning process from data provided by the OLAmobile company and with high volume and velocity characteristics (big data streams). The algorithm will consider several innovative and differentiating elements when compared with the state of the art: use of behavioral variables (e.g. user history of navigation) and post-subscription (e.g. most consumed products and categories) that are collected via a new Application Programming Interface (API) to be developed by OLAmobile; and use of advanced Machine Learning (e.g. Neural Networks) and Modern Optimization (e.g. Evolutionary Algorithms). The Modern Optimization engine will be a key component of the algorithm, automatically optimizing several objectives (i.e. interests of those involved in this business), searching for relevant input variables and also adjusting the machine learning model to the data in real-time (online training). As interesting results will come up at UMinho, the algorithm will be adapted and implemented for the OLAmobile technology, with the goal of validating the algorithm in an industrial environment (TRL 5), with its operation in a real environment. When the algorithm of the prototype is complete, it will be put into operational use (TRL 7), reaching a new technology for full use by OLAmobile and with high impact on its Mobile Performance Marketing industry.
The new technology, researched and developed in this project, will add an innovative intelligence layer to the Mobile Performance Marketing area. Therefore, an increase of profitability is expected for the OLAmobile company and its associated agents, with a growth of advertisers, publishers, mobile content creators and even users. Moreover, such technology will aid OLAmobile to enlarge its mobile market presence, particularly in the Asia-Pacific region (e.g. Indonesia, Malaysia, China). Furthermore, this project will have several other outcomes (e.g. two PhD thesis, two MSc thesis, a website, an international patent application, publication of international scientific articles) that will be widely disseminated (e.g., Web, social networks and other media).