Dan Zigmond, Technical Lead for Google TV Ads, had a post in the Official Google Blog over the weekend describing how Google analyzes the data from the TV ads they run. Using data from set-top boxes, they can measure for each ad how many viewers stay on-channel versus those who switch over. This information is used to determine each ad's effectiveness compared to a statistical model. It's explained in this video:
R isn't explicitly mentioned in the post or video, but given that R is widely used at Google, and that the charts look like R and that this is a kind of modeling R is particularly suited to, it's a safe bet R was used. I tried contacting Dan Zigmond to confirm, but no response yet. Dan, if you're reading, please let us know!
Official Google Blog: Tuning in to TV data. (Thanks to reader DB for the tip.)
Comments