In general, people tend to confuse means and ends.
Programmatic TV/radio is a good example of it in our industry. There's an assumption that because programmatic is growing in digital, it must therefore be the future for TV and radio as well. But to believe that is to confuse means and ends.
Programmatic has grown by being a solution to a particular problem that is specific to digital, namely a vast array of fragmented inventory. More specifically: how to monetize this fragmented inventory effectively; how to do so efficiently when a buyer can't realistically contact each site, and how to assess the cumulative effect when you're done. A programmatic approach helps alleviate those issues, in that it reduces the number of sales points drastically whilst still monetizing multiple fragmented sites.
In other words, programmatic was and is primarily a solution to an inventory management problem, wherein there was too much unsold inventory representing small numbers of people, scattered across too many different places.
This is not a problem that exists in TV or radio in Canada. TV and radio does not have lots (or any) unsold inventory; in Canada at least, inventory typically sells out. TV and radio don't have lots of inventory representing small numbers of people; even a small show on TV or radio still brings in thousands of people on any given minute. TV and radio don't have thousands of sales points for a buyer: instead, they have a handful, and if anything that number is shrinking as the industry consolidates further.
So to assume that programmatic TV or radio should happen is to confuse means and ends. It's to assume that 'programmatic' is a generic 'good thing', rather than a solution to a particular problem. It ignores some of the significant challenges seen so far, including transparency and the number of middlemen taking a cut on the way through the process. Even worse, it's to fall prey to some clever spin wherein the solution to a problem has been turned into an all-purpose benefit.
That's not to say that increased automation won't come to TV and radio. My assumption is that it will, in some shape or form, in the next 1-2 years. The key to determine that shape or form is to determine what problems TV may have, and to see what aspects of programmatic (deliberately loosely applied!) could help. For example, one area that springs to mind is the need for TV stations to maximize yield on all inventory, not just ads that are in the shows most watched by broad buying demos. To do that, buyers and sellers need a mechanism to determine value quickly on audiences that don't already have market prices baked in. Doing that manually takes time and effort; doing it algorithmically could be quick and easy.
But let's make sure that, as we look for ways to improve our industry, we spend some time distinguishing the means and ends first. Let's fit the right solutions to the right problems, and build what's best for TV and radio.