Programmatic, or automated buying and selling, has been developed primarily for digital media, where there are innumerable "publishers" looking to sell ads on their websites, many of whom have no national sales representation and may be overlooked by a human buyer. The operating principle, as we understand it, is that each publisher inputs ad rates, with discount structures if possible, as well as relevant "audience" data. The buyer's "trading desk" then scans all of the avails to come up with the best way to reach the specified target group at the most cost efficient rate.It's more complicated than that, of course, and various algorithms are employed to take account of the advertiser's timing needs, demographic or other target weights, content types, etc. But, in the main, programmatic is driven by cost efficiency, at least that's our take.The question is whether this approach can be utilized for buying TV? Will it yield the promised undreamed of targeting efficiencies and do away with TV's outmoded age/sex buying criteria? According to programmatic enthusiasts, the answer to this question is a resounding "Yes."The networks are playing this particular game with the deck stacked against them. To begin with, their average minute currency is based on those sets that were tuned in to a program while commercials were aired. Any time the commercials were zapped—mainly by DVR users—such audiences were deleted. In contrast, digital video ad sellers have been charging CPMs comparable to network prime and much higher than the average for all TV dayparts and network types, even though only 50% of the ads served will appear on the user's screen for only 1-2 seconds, and a much smaller percentage are viewable in their entirety. Hence a 30-second primetime network CPM for adults aged 18-49, which averaged $41 in 2014, gives all of its measured viewers a full opportunity to see the ad. However, an untargeted digital video ad, which may be a 15-second or a 30-second message, comes in at a CPM norm of $52 to start with and this rises to something like $130 or more when those who can't see the ad from start to finish are deleted. For some reason, the networks have been very slow to point out such distinctions to time buyers and advertisers. Why? To support their case, the programmatic side claims that for the first time, the adoption of "big data" TV set top box ratings combined with ascribed sales or product use data from third party sources will enable buyers to buy audiences, not GRPs, though they admit that Nielsen ratings will probably remain the currency for most buys. In effect, big data findings on a show-by-show basis merged with product purchase information will create market value indices that can be melded with the Nielsen ratings to tell the system how valuable the audience is. Toss in costs and the trading desk can sift through all of the avails posted by the sellers and come up with an optimally efficient buy, often including small audience networks or channels and/or programs the buyer may not have been aware of. How does this idea strike you?