Coffee with Michael and Tom is a regular series of videos where Dativa's CEO Michael, in California, and CTO Tom, in England, discuss Dativa's work with TV datasets. In the first episode, they begin by drawing parallels between creating the perfect cup of coffee by blending various beans together and creating the perfect granular TV data by blending various datasets. Obtaining this granular TV data is not about starting with the ideal dataset; the trick is in the blending. Dativa is currently releasing its Best Practices white paper, which discusses the types of datasets typically used in TV data and the best practices for combining them.
Michael and Tom then ponder as to why exactly this issue of blending and combining different datasets has become such a hot topic right now, and to what extent this interest is being driven by the demand and supply sides within the TV advertising industry. Tom predicts that people in the industry are shortly going to stop talking about how enormous datasets are and will instead focus on the expertise of their in-house analysts to identify which are the right datasets to use for any given problem. The process of taking these big datasets and dealing with them in-house is one that Dativa is seeing plenty of recently and we predict that this is a trend which will, and equally importantly certainly should, be happening a lot more.
In 2020, everyone is building their own data platforms using their own first-party data, and this process is dominated by digital. Those companies that don't already have a D2C service are in the process of launching one, and what they all find is they have to cope with gigabytes of new data on a daily basis. Tom states that what really excites him is not combining STB and Smart TV datasets but combining either one with streaming data.
A lot of the streaming data is connected TV data, and this is a booming area right now. With linear, Michael sees that the skews and biases tend to be driven by demographics and not by the TV provider. In contrast, CTV data is subject to platform bias, including what apps are available and how much the particular app emphasizes user experience.