PrediCube makes online advertising more effective while respecting consumers' privacy
27 February 2015
Belgian tech starter PrediCube, a spin-off from digital research center iMinds and University of Antwerp, uses big data analytics to make sure consumers get to see those ads that are truly of interest to them, thereby increasing click-through rates up to 300%, while putting its unique 'privacy by design' strategy center stage.
Online advertising is big business, but spamming consumers with random ads has proven not to be the best way to optimize advertising returns: people only click on ads that are relevant to them. Hence, media companies and advertisers have been exploring targeted advertising strategies to boost online ads’ click-through rates (and revenues) based on an analysis of people’s Internet surfing patterns. An approach that conflicts with consumers’ stringent privacy concerns? Not anymore, thanks to Belgian tech starter PrediCube – a spin-off from digital research center iMinds and University of Antwerp, and supported by the Start it @KBC incubator. PrediCube uses big data analytics to make sure consumers get to see those ads that are truly of interest to them, thereby increasing click-through rates up to 300%, while putting its unique ‘privacy by design’ strategy center stage.
Analyzing and predicting consumers’ online behavior to increase click-through rates up to 300%
Online advertising is big business. In Europe alone, online ad spending topped €27 Billion in 2013 (a YoY increase of 11,9%). Yet, while this big market potential offers a great deal of opportunities, the shift to online advertising also comes with a number of important concerns. The authors of the book Het nieuwe TV-kijken found, for instance, that 70% of every euro that is redirected from print to online advertising currently floats out of the local economy – right into the hands of a few big international players such as Google and Facebook.
Trying to counter that drain of resources and valuable consumer data, PrediCube now brings to market a solution that can be used by local media advertising companies to predict consumers’ interest in specific ads – based on an analysis of their online behavior. Result: targeted online advertising campaigns that are much more efficient and generate higher revenues.
“Using PrediCube, we have been able to increase targeted online ads’ click-through rates up to 300%. In other words, we are now able to match ads with the right consumer profiles up to 3 times more accurately. Moreover, we want to significantly increase the inventory’s volume based on socio-demographic criteria (such as age and gender). Thanks to PrediCube we will be able to do this in the very near future. It goes without saying that this approach will positively impact our business and product offering,” says Philippe Degueldre, director business intelligence at Pebble Media, managing online advertising for 80 premium websites – such as VRT, Telenet, RTBF, Viacom, Elle and LinkedIn.
Strong focus on ‘privacy by design’
Dealing adequately with consumers’ privacy concerns is a major focus area for the PrediCube team; hence they are investing a great deal of effort in their ‘privacy by design’ approach.
“PrediCube works by means of cookies,” explains Prof. dr. David Martens, co-founder of PrediCube and Assistant Professor at the Faculty of Applied Economics, University of Antwerp. “First of all, web pages that use the PrediCube customer behavior prediction tool will ask users’ explicit consent to use those cookies. If the cookies are not accepted, no behavior tracking will take place.”
“Secondly, a number of privacy safeguards have been put in place,” David Martens continues. “Users are automatically ‘forgotten’ after 30 days, their online behavior is only tracked on premium web pages – not across the whole of the Internet – and their data is never sold to other parties.”
PrediCube: bringing together the best in research and entrepreneurship
PrediCube builds on the outcomes of the DiDaM project, a collaborative research effort under the banner of iMinds Media. DiDaM investigated ways of analyzing media users’ Internet sessions in real-time and identifying patterns to help advertisers integrate relevant ads into web pages (also in real-time). Objective: providing consumers with just those ads that are relevant to them. DiDaM’s research findings and the expertise from the Applied Data Mining research center of the University of Antwerp together laid the foundation of PrediCube. The PrediCube team can also count on the support of the Start it @KBC incubator – providing them with business guidance and office space.
PrediCube is a new tech startup that uses advanced big data technology to predict which online users are interested in a product, allowing targeted advertising on premium websites. Or how a spinoff company of University of Antwerp and iMinds is ready to go head to head with Facebook and Google to compete for online advertising budgets. PrediCube results from an iMinds Media project that ended in 2014, investigating the potential of data for improved advertising (together with partners Concentra, Pebble Media, AdHese and KU Leuven). PrediCube has been co-founded by prof. David Martens, who heads the Applied Data Mining research group at the University of Antwerp (faculty of Applied Economic Sciences), and whose research focuses on the development and application of data mining algorithms. PrediCube was founded in October 2014, and is part of the Start it @KBC incubator. Current customers include Batibouw, Engels Ramen, Verandaland and Triple Living.
Prof. Dr. David Martens
+32 476 91 87 09