Re: Barriers to purchase @ Eric Olson
Nielsen's method is a combination of boxes, surveys, and reports from content providers. They also can figure out how many people are watching in a time-shifted manner. Is it perfect? No.
But the numbers aren't what matters, it's the behavior of the consumer. Those watching TV will have a certain engagement and conversion rate, modeled over decades and actually compared to consumer habits. The internet provides very little of that, for a number of reasons.
For one, the population of those who stream television through a service tend to be younger, more tech savvy individuals. They also are ad-adverse and will likely leave the room to grab something for that 30 second spot. Or YouTube often offers a "skip" button 5-10 seconds in. In general, your count of eyeballs online is likely to be over-inflated and have a much lower conversion rate. What's the point of having perfect numbers when you can't actually see how many people are watching the ad.
TV is usually more communal, and even now is a center of a home where 2 or more people will gather. Short of a handful of HTPCs that are setup by tech geeks, most internet content is streamed to a device that is watched by one person. So even if you can exactly say that 2.5 million views of that show were done online, you have to assume it's only 2.5 million (or less) viewers. For tradition delivery, 2.5 million households (views) on a Tuesday night could be 5 million or even 7.5 million people, and you can survey that information easily. More importantly, a much larger number of them are used to being swayed by ads and accept it, increasing engagement and conversion rates.
Finally, if you are seeing ads that are not applicable to you, it's probably because you are watching something outside your demographic. If you are in the US and watch CBS during primetime, you will see a lot of things that skew old and "comfortable". That's CBS's core demographic, the silver-haired Boomers who like shows that are procedural acronyms. Watch Mad Men on AMC, and you'll see a lot of luxury car ads and upscale beverages, because the demo that watches are younger, affluent, and more likely to drink alcohol that looks expensive, but not be learned enough to know what good alcohol is (that last bit is editorializing on my part, but I digress).
The point is that the internet, specifically streamed TV and movie content, is a very new ballgame. The models are incomplete, the numbers are suspect, and the viewers are less likely to engage with the ads in the first place. So not only can free-to-air TV do things that internet-streamed TV cannot do, it's cheaper, more targeted, and more likely to have a better return on investment. This could change in the next 10 years, but I think (hope) there are more profound content delivery changes on the horizon that will make this conversation moot.