Meet Recombee at IBC 2023 Amsterdam · 15–18 September 2023 · Read More

Complete Editorial Content Automatization for Multiple Platforms

Prima Group is one of the strongest media companies in the Czech Republic with average monthly viewership consisting of nearly 100% of the Czech economically active population. Prima Group broadcasts seven free-to-air television stations and has a wide portfolio of websites.

Recombee's smart content recommendations helped Prima to completely automate their editorial content for multiple platforms (desktop, mobile, hbbtv) and applied recommendations to 55+ different places on their platform (scenarios).

Recombee’s high-tech recommender engine brought a significant increase in the volume of advertising consumed online.

15%
Increase in click-through rate
3.5%
Increase in the number of views

Situation

  • A very large volume and wide variety of video content available across multiple platforms.
  • Content recommendations made manually predominantly by editorial staff.
  • Desire to improve quality of personalization for millions of monthly visitors.

Requirements

  • Personalized content recommendations for every individual visitor.
  • Personalization on multiple platforms (desktop, mobile, hbbtv).
  • Plot continuity of the recommended content.

Solution

  • More than 55 different unique scenarios on various platforms
  • Fully personalized homepage using 5 competing scenarios
  • Sophistically tailored algorithms to meet the specific requirements for different scenarios

Benefits & Results

  • 15% increase in click-through rate on the homepage scenarios
  • 3.5% increase in the number of views on the homepage scenarios
  • A rise of profit from the advertisement shown across the website including the homepage

Scenarios

Homepage recommended

Items to User scenario used for recommending similar content to homepage visitors. Utilizing default custom model developed specifically for FTV Prima brings higher visitor’s engagement right from the beginning.

Selected For You

Another Item to User scenario used at the bottom of the homepage. This highly performing custom model is utilizing Recombee’s unique collaborative filtering models analyzing visitors’ behavior online to recommend content that will most likely spark the interest of viewers.

Related Series

Items to Item scenario set up specifically for series recommendations. Alongside content and collaborative filtering models, the advanced business rules (e.g. genre similarity booster) have been applied to recommend similar shows by genres such as romance, comedy, or medical drama.

Similar Movies

Another Items to Item scenario using a custom-developed model with an increased level of randomization (shuffling) to provide recommendations of similar movies for mobile device users.

“Our cooperation with Recombee is very fruitful. We are having periodic calls where we are discussing possibilities for adjusting and improving current models to perform better for our use-case and also looking for new places for the deployment of personalized recommendations.

So far, Recombee helped us to increase our CTR on the homepage by 15% and boost video views by 3.5 %. At the moment we discuss recommendations from linear TV to video on demand archive, which can bring our cooperation to even higher level. Highly valued partnership!“

Jan Lajka
Head Of Data & Programmatic, FTV Prima
FTV Prima

About FTV Prima

FTV Prima was established in 1993 as the first commercial Czech TV station. The company currently broadcasts 8 TV channels and 6 radio stations and has a broad portfolio of websites and printed magazines. It is a leader in the Czech market in HbbTV.

Their most popular website iprima.cz is visited by millions of viewer per month, consuming wide range of content, such as daily news, tv series, movies or documentaries.