The decline in newspaper sales has led to a change in the business model of the print media sector. The large publishing groups have redoubled their efforts to offer the best digital media. Pentaquark started a collaborative project with one of the most important publishing groups in the country. The main objectives were to increase the number of subscribers to its newspapers in digital version and to develop a score model to estimate the propensity of each client to purchase one of the products promoted on the newspaper's website. To do this, Pentaquark developed and implemented a Leads estimation model and carried out an automated collection of relevant information from users (Telephone, mail, web registration,). Simultaneously, our staff developed a model based on the segmentation and profiling of users through logistic regression, which determined the propensity of users to contract a subscription. With the results obtained, personalized offers are made to each user, carrying out more precise Marketing actions and with higher return rates.

1054/5000 After having obtained very favorable results, our client decided to extend the collaboration with our company to carry out the optimization of the negotiation process of the fees imposed on third parties for the publication of “Advertising banners”. The objectives were based on providing the publishing group with the ability to predict the number of users who would access an ad and allowing the advertiser to use the newspaper's website as a distribution channel for its products in exchange for giving a margin to the publishing group. Pentaquark developed a “Click-through rate” model that determined the number of users who accessed the ad and, later, a Machine Learning algorithm would be implemented to determine the different variables that lead to a greater number of accesses to the ad. With the results obtained, a renegotiation process is carried out with the advertisers to optimize the economic performance of the website itself, allowing the elimination of traditional auction valuation methods.