There are a lot of conversations going on in media right now. While Facebook eats up targeted dollars from advertisers, it continues its pitch against TV. Only people who are smart know when selling attention, brands care about time. And yet, false comparisons between digital and TV persist.Some brands buy into the innovations in social, some brands opt-out of the hype and who the hell knows whats working?Consider last week:Nielsen ingested Facebook, Hulu and YouTube into its digital metrics. Not its TV metrics. Meanwhile in ratings land, there is a strong case made by Inscape’s EVP as to why Comscore and Nielsen will always be off and why using smart TV data, (like that of VIZIO) solves the issues legacy systems can’t. And David Bloom wonders if TAM Can Keep Up With Media’s Fracturing Reality. When it comes to live linear TV, there is very little evidence that a lower ad load increases performance of TV ads, yet networks are testing it because there is definitely evidence that scarcity brings up prices when audience numbers are down. And while the odds of a completion rate should go up when an ad is 6 seconds over 30, that’s not a given, either. For one, lots of 6 second ads may be more sickening than a few 30s. For two, in general on live linear, when people leave ads on, they generally stay on. EG-- Saturday Night Live advertisers get played all the way through 85% of the time, according to TV attention metrics  from iSpot.tvTeens Choose DifferentlyWhile the Teen Choice Awards got more people emotionally charged on social than Game of Thrones this week, and ratings were up YoY. Yet data from Inscape’s smart TV population had that show falling outside the top 100 for the week. Meaning it was outpaced by under-hyped gardening shows and re-runs. That speaks to the point that there are many places and many ways to experience TV content these days and that and content engagement from TV-programming doesn’t always equal live-linear eyeballs.Canvs emotional analysis on 263k emotional reactions to 2017 Teen Choice Awards.