The conversation between me and the interviewer (right, I am looking for another opportunity) yesterday made me think a lot about how advertisers look at which media and how much to invest, and how the marketing as a whole still needs a lot of improvement in terms of efficiency.
Like on the online advertising side, you have google, yahoo and microsoft which make up the head of the long tail curve. They utilized their resources (machines, technology, people) just to figure out what the user is really interested, what he is going to click next and also of course to grab as much market share as possible in the online world. Machine learning, collaborative filter, etc, etc wowed a lot of us. The competition is so intense that you almost feel unable to breathe for even a second!
On the other side, when advertisers look at the budgets in their pocket, internet is just one channel for them to broadcast their products, their brand. They probably rely on the analysis done by firms that do marketing mix analysis. Those people seem to come from complete different background with those in the online world. They often have training in economics and marketing. The model they use to analyze lift from different media and recommend mix, is more or less in the same level of linear regression.
So think about this, it is those "simple" models that recommend how much money actually going into this industry. It may not be very efficient as a very important starting point. And the internet firms use all the "fancy" stuff to make sure those money got spend efficiently. How ironic this is! Of course, there is nothing wrong using simple models. What shocked me is really the contrast!