RTL Nederland analyses social media buzz to optimise its product offering

Using IBM SPSS Predictive Analytics software to gain insight into viewers’ preferences

Published on 04-May-2011

"RTL Nederland constantly incorporates much of the viewer feedback to optimise its product offering. When approaching the final episodes of the X Factor, the online buzz about the candidates even increased by about 400 percent, providing a very rich source of information on viewer likes and dislikes. IBM analytics helps us to better understand audience needs and preferences, and hence increase viewer involvement." - Emilie van den Berge, senior Research & Intelligence project leader, RTL Nederland

Customer:
RTL Nederland

Industry:
Media & Entertainment

Deployment country:
Netherlands

Solution:
BA - Business Intelligence, Smarter Planet, Information On Demand

Smarter Planet:
Smarter Media

IBM Business Partner:
InSites Consulting

Overview

RTL Nederland is a trend-setting multimedia and entertainment company with a leading position in the Dutch broadcasting market. It owns five television channels, two radio channels and an extended internet network. Twenty-four hours a day, seven days a week, the TV channels RTL 4, RTL 5, RTL 7 and RTL 8 and the digital channel RTL Lounge amuse and inform viewers with programmes such as news and show-business, soap operas, international soccer, and a diverse array of movies and TV series.

Business need:
RTL Nederland aimed to evaluate various television programmes in the Dutch market and increase viewer involvement by making use of online conversations. The company needed a way to analyze, interpret and successfully respond to audience feedback from social media sources.

Solution:
RTL Nederland worked with InSites Consulting to capture viewer opinions from user-generated comments on social media and other online buzz by using IBM predictive analytics software. This helps RTL Nederland to better understand audience needs and preferences, and hence increase viewer involvement. The insight obtained on viewer likes and dislikes allows RTL Nederland to optimise its product offering.

Benefits:
Analyzed the sentiment of over 71,000 online conversations about X Factor, providing RTL Nederland with a powerful tool to measure attitudes indirectly and quickly adapt the program accordingly. Captures unstructured data automatically from the web with sophisticated text analytics technology. Approaching the final episodes of the reality competition shows, online buzz on the program increased by about 400 percent, which provided a very rich source of information about viewer opinions.

Case Study

To read a Dutch version of this case study, click here.

Instrumented: Captures unstructured data automatically from the Web with sophisticated text analytics technology.

Interconnected: Approaching the final episodes of the reality competition shows, the online buzz in one instance increased by about 400 percent, providing a very rich source of information to RTL.

Intelligent: Aggregates data on viewer preferences based on more than 71,000 online conversations, allowing RTL to enhance decision-making and optimise its product offering based on reliable user feedback. Increases viewer engagement as well as RTL’s ability to monitor its viewing community, providing a rich source of information on viewer likes and dislikes.


RTL Nederland is a trend-setting multimedia and entertainment company with a leading position in the Dutch broadcasting market. It owns five television channels, two radio channels and an extended internet network. Twenty-four hours a day, seven days a week, the TV channels RTL 4, RTL 5, RTL 7 and RTL 8 and the digital channel RTL Lounge amuse and inform viewers with programmes such as news and show-business, soap operas, international soccer, and a diverse array of movies and TV series.

The main objective of RTL Nederland is to attract, retain and serve its viewers by offering high-quality and distinctive programmes. The company therefore places great emphasis on evaluating programmes on a continuous basis. Research helps RTL Nederland stay in close contact with the audience’s opinions and wishes, and helps to optimise the programmes that are broadcast.

Information needs
The company is constantly looking for new and better ways to perform this type of research, and realised that two of its most popular series, So You Think You Can Dance and X Factor, offered a good opportunity to make use of viewers’ opinions to improve the quality of the programming.

“The scope of our research for the series was twofold: on the one hand to gather general findings that can be used for future broadcasting of the show, and on the other hand to evaluate specific aspects in order to make adjustments during current broadcasting,” explains Emilie van den Berge, senior Research & Intelligence project leader at RTL Nederland. “Important research topics were the jury, the choice of music, the themes of the Live Shows, the programme’s Internet pages, the presenters, the scenery and of course the contestants themselves. Since the auditions and Boot Camp are recorded before broadcasting, potential adjustments can only be made during the Live Shows.”

A new approach
In addition to relying on viewer ratings and traditional market research methods such as surveys and panel discussions, RTL Nederland wanted to take advantage of other data to evaluate its programmes, such as comments that viewers were making about its shows on social media sites. To achieve this, the company adopted a new methodology known as ‘netnography’, which has been pioneered by market research consultancy, InSites Consulting.

“Netnography, or social analytics, provides a method of mining the unstructured data that is held in user-generated content that is freely available from social media sites such as Facebook, Twitter, YouTube and hundreds of other blogs, discussion forums and web sites,” says Niels Schillewaert, co-founder and managing partner of InSites Consulting. “The approach uses IBM SPSS Predictive Analytics software to analyse thousands of comments, identify important themes and patterns, and compare them to find out which are the most important – the topics that have the most ‘online buzz’.”

Tuning in to viewers’ opinions
RTL Nederland began its trial of social analytics with So You Think You Can Dance (SYTYCD), a TV talent show in which candidates compete in a series of dance contests, are judged by a panel of experts, and then face a public vote to determine which of them are good enough to go through to the subsequent rounds.

A combined team from RTL Nederland’s Research & Intelligence division and InSites Consulting started by selecting the most appropriate social media sources, choosing both popular sites with large numbers of visitors (such as hyves.nl and www.forum.fok.nl), and smaller, more specialised sites where the most dedicated fans tended to congregate.

“It’s really important to select the right sample of sources at the start of the project, to limit the amount of ‘white noise’ that can reduce the relevance of your findings,” explains Niels Schillewaert. “In this case, one of the main challenges was that So You Think You Can Dance is an international show, and different countries have their own versions of it. We were only interested in the audience’s reaction to the Dutch version of the show, so we decided to limit our sources to comments written in the Dutch language and posted on sites with a .nl domain name.”

Once the sources had been selected and a set of SYTYCD-specific keywords had been identified, the team began extracting the data from the websites – an automated process that not only captures the text itself, but also gathers as much metadata as possible about the user who posted it and the number of views or approval ratings that the post has received from other users.

Using predictive analytics
In total, this process collected 14,000 comments about SYTYCD, which were then subjected to detailed analysis using IBM SPSS Modeler. The software uses a library of predefined words to find common word patterns, providing insight into the main topics that viewers were discussing and identifying which aspects of the show were creating the most buzz. The team also used a sophisticated ‘sentiment analysis’ to identify emotional vocabulary and decide whether this buzz was primarily positive or negative.

“We never directly interviewed participants about the television shows, but conducted a sentiment analysis afterwards to have an idea of how good or bad they considered each program to be,” explains Emilie van den Berge. “For instance, we decided to change the voting procedure in the middle of the live shows of SYTYCD. Sentiment analysis provided a powerful tool to measure attitudes indirectly after the shows. It demonstrated an increase in the positive buzz, indicating the viewers liked the adaptation of the programme format.”

Identifying the X Factor
Impressed by the results of the project, RTL Nederland decided to make IBM SPSS Predictive Analytics a central part of its research toolset for its next major live talent show, X Factor. The show uses a similar model to SYTYCD, where contestants are rated on their performance as singers across a variety of musical genres.

“The X Factor live shows were broadcast on Friday evenings, and we collected the social media data on the following Monday,” comments Emilie van den Berge. “We reported our findings to the programme’s production team on Wednesday, which gave them two days to incorporate the feedback into the next live show. This meant we were able to adapt the show to viewers’ wishes from week to week.”

Listening to the buzz
Over the course of the series, the team collected more than 70,000 comments from social media sites, and found that viewers had opinions on almost every aspect of this extremely popular show.

“They were talking about everything from the clothes the presenters were wearing to the comments of the judges, the choreography of the dancers, the performance of the contestants, and the production of the show itself,” says Emilie van den Berge. “IBM provided us with a solution capable of analysing these very large volumes of data to gain insight in viewer feedback that we were able to put to productive use in improving the show.”

Real results
The analysis not only showed which contestants had the most positive and negative buzz – it also highlighted the contestants who provoked no opinion at all. To avoid the risk of viewers becoming bored when these contestants were on screen, the producers used a number of strategies to raise their profiles. In one case, a contestant was given a complete makeover – a new style of clothes, more fashionable glasses, and more striking songs to sing and dance to. In the end, he became one of the most popular acts on the show, and even reached the series’ final.

“We were also able to take a number of viewers’ suggestions on board. For example, each live show had a different musical theme: ‘Motown’, ‘The Nineties’, ‘Go Dutch’, and so on. We saw a high level of buzz about both the themes and the individual songs that were chosen: many viewers felt that they should have greater input into this area of the show. As a result, before the final live show of the series, we introduced a voting system on the X Factor web site that allowed viewers to choose their favourite songs for each finalist. This immediately led to a much more positive buzz on the topic of music choice.

She adds: “There was also a lot of buzz around the making of the show itself, and a lot of interest in what went on behind the scenes. To satisfy the viewers’ curiosity, we gave each of the contestants a camera and asked them to make a short video about an average day on the X Factor. These videos were made available on the web site, and again, we received a high level of positive buzz as a result.”

Increasing viewer involvement
The project demonstrated that analysing and deploying feedback obtained from social media provides RTL Nederland with useful input to enhance decision making.

“RTL Nederland constantly incorporates much of the viewer feedback to optimise its product offering,” says Emilie van den Berge. “When approaching the final episodes of the X Factor, the online buzz about the candidates even increased by about 400 percent, providing a very rich source of information on viewer likes and dislikes. IBM predictive analytics help us to better understand audience needs and preferences, and hence increase viewer involvement.”

“Social analytics has the potential to be a very valuable tool for research in broadcasting and other industries, because it allows the data to speak for itself,” concludes Emilie van den Berge. “At RTL Nederland, we are excited to be participating in the development of this new science, and we hope to continue to work with InSites Consulting and IBM to evolve the methodology further. By listening to our viewers, we give them the power to contribute to the creative process – helping us to produce better, more compelling television for our audiences.”

About IBM Business Analytics
IBM Business Analytics software delivers complete, consistent and accurate information that decision-makers trust to improve business performance. A comprehensive portfolio of business intelligence, predictive analytics, financial performance and strategy management, and analytic applications provides clear, immediate and actionable insights into current performance and the ability to predict future outcomes.

Combined with rich industry solutions, proven practices and professional services, organisations of every size can drive the highest productivity, confidently automate decisions and deliver better results.

Products and services used

IBM products and services that were used in this case study.

Software:
SPSS Modeler, SPSS Text Analytics

Legal Information

© Copyright IBM Corporation 2011. IBM Nederland hoofdkantoor, Johan Huizingalaan, 7651066 VH, Amsterdam. Produced in The Netherlands, April 2011. All Rights Reserved. IBM, the IBM logo and ibm.com are trademarks of International Business Machines Corporation, registered in many jurisdictions worldwide. A current list of other IBM trademarks is available on the Web at “Copyright and trademark information” at ibm.com/legal/copytrade.shtml. SPSS is a trademark of SPSS, Inc., an IBM Company, registered in many jurisdictions worldwide. Other company, product or service names may be trademarks, or service marks of others. References in this publication to IBM products, programs or services do not imply that IBM intends to make these available in all countries in which IBM operates. Any reference to an IBM product, program or service is not intended to imply that only IBM’s product, program or service may be used. Any functionally equivalent product, program or service may be used instead. All customer examples cited represent how some customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics will vary depending on individual customer configurations and conditions. IBM hardware products are manufactured from new parts, or new and used parts. In some cases, the hardware product may not be new and may have been previously installed. Regardless, IBM warranty terms apply. This publication is for general guidance only. Photographs may show design models.