To date, TV has mostly been measured through a baseline and lift model. It is a framework that works well for linear TV, since many people watch the same program and advertisements at the same time. As such, even if only a small fraction of viewers responded to the ad, the lift is still visible and noticeable above the baseline.
This traditional framework is increasingly effective, owing to the availability of more data and improved analytics. It was most recently reinforced with data from smart TV providers. This allows advertisers to approach TV measurement as if it were a digital campaign, including leveraging IP-level data in a closed-loop attribution model and calculating delayed lift or reach and frequency with deadly accuracy.
Such measurements, done in near real time and available in beautiful online dashboards, is what TV advertisers rightfully demand. They want to drive their TV campaigns as if they were digital. As a whole, the measurement of TV advertising has made significant progress in the last 10 years.
TV-viewing behavior, however, is changing rapidly and undermines recent accomplishments. Consumers want their content on-demand, whether delivered through over-the-top (OTT) via Hulu, for example, through connected TV (CTV) with a Roku device or through video on-demand, such as HGTV or AMC. While on-demand is thought to be mostly viewed on mobile and tablet devices, the big screen in the living room is still the preferred consumer experience. Therefore, measuring response from on-demand TV cannot be done effectively by tracking “clicks on the screen” or a typical digital-response model.