Introducing our TV Analytics Dashboard 2.0

TV attribution is hard as you do not have any direct tie (e.g. a click) between the ad and the viewer interaction. Due to the direct tie, digital campaigns can be tracked quite easily on a user basis. For TV attribution, you need to build your own model based on the airing time of the commercial and the web/app traffic of the TV advertiser (here is a general introduction on how it works).

The basic principle of calculating a baseline and attributing the uplift to a particular airing is quite simple. The hard part is to automate what the human eye can do with intuition, i.e. interpreting a visual traffic flow chart by determining a variable baseline and a variable attribution window for each airing. We have seen many clients using simple automated models with fixed attribution times and fixed baselines which oftentimes recommend the complete opposite of what they should.

Today we did a massive update to our attribution algorithm, the next milestone on our path to providing the best TV attribution possible. It automatically detects anomalies in web traffic (let’s say there is a mention on TV which sends thousands of people to the website within minutes), understand fluctuations in uplifts (let’s say some viewers visit your website right after your ad starts, some just at the end of the ad, creating two traffic peaks for one airing), and handles fast rises and falls in general website traffic beautifully. This allows us to better determine the variable baseline and variable attribution window for each airing. Below you’ll find a picture illustrating this.


The beauty of the new algorithm lies in the fact that it does not build upon “typical” patterns, making it one of the industries’ most accurate solutions. You might have heard of many typical patterns such as:
During primetime, we typically have double the traffic than during the afternoon. We typically sell double the amount on the weekend than on an average weekday. On average, TV viewers visit our website up to 5 minutes after the airing.

While all of these statements might be true, going with typical patterns does not cut it when it comes to attributing website/app traffic and conversion down to each single airing. Why? Because you might have heard these statements as well:
Yes, Monday was a holiday, that’s why traffic was slow. We kicked off our massive online campaign last week, that really boosted traffic. We sent out our newsletter to 3 million people at 11 AM yesterday.
If you rely on typical patterns, you need to (most likely manually) account for all atypical events.

What does combining our real-time detections with our updated attribution model mean for you? Real-time optimization of your TV campaign with industry-leading accuacry you can rely upon. Reach out to us to schedule a demo.

Your contact for the US & Canada: (other countries)
Victor Castello