Channel 4 were selling advertising space on their video on demand (VoD) platform in the same way that they sold display advertising – as a highly commoditised product with a relatively low cost per million (CPM). As the incumbent data science agency, Realise Unlimited was tasked with helping Channel 4 to understand targeting groups to better serve ad inventory.
VoD is no different from linear (broadcast) television where different demographic groups are sold at a premium, therefore there was a significant opportunity to sell VoD inventory at a higher CPM. In order to realise this opportunity, it was important to be able to understand the market, size the potential opportunity and most importantly explore what data were available to provide a new source of inventory that could be sold at a higher CPM.
We developed a suite of predictive models to identify registered viewers with a high probability of being ABC1 (social grade), house boss or house boss with kids. One of the learnings from these initial models was how predictive engagement with different types of content was for identifying different demographic groups. As well as focusing on registered viewers, additional demographic targeting models were built for anonymous (cookie based) inventory. This was a significant win for the business because it released a large amount of additional inventory that could be sold at higher rates. And, to ensure the validity of the new inventory, the models and data were audited by PWC prior to the new advertising products being taken to market.
The ability to assign demographic targeting groups to known and unknown viewers to determine the right ad inventory to serve. 52% of the known base were identified as ABC1.
There was significant commercial benefit resulting in £10m profit in year one.