I write about privacy a lot. I do it because it is an issue that I care about a lot, and because it is having more and more impact on our industry.
Internet-like, data-driven advertising is now expanding beyond the PC to new classes of connected, digital devices, from smartphones to tablets to televisions. Many within our industry are watching this expansion, given the potential to transform how tens of billions of ad dollars are spent each year. Many outside our industry are watching as well, wondering whether we will do a better job handling privacy on smartphones and smart TVs than we did with PCs.
I wonder, too, but also have the advantage of being able to do something about it.
We had the ability to deliver fully addressable ads based on prior Web browsing history starting in 1996 and 1997 through the state-of-the-art ad servers of that day. But it took at least 10 years until it represented even 5% of total online ad spend. Clearly, just because something is possible and theoretically valuable doesn't mean that the market will buy it and use it. These things take time.
1-to-1 advertising has never really caught on in practice on the Web. (Yes, many e-commerce merchants like Amazon use 1-to-1 constantly, but that is not how the vast majority of online ads are delivered today.) That is why, when it comes to tailoring ads on TV, where I spend my time these days, I am a big proponent of focusing first on using data (such as Zip code-level viewing, census and sales data) to create better broad, anonymous targeting segmentations, rather than trying to created granular profiles of individual viewers or households. Here are my reasons:
Today, even if we wanted to, only a small percentage of TV households in the U.S. can be reached with addressable advertising. Kudos to folks like Cablevision for leading the way and building an extraordinary targeted ad-delivery platform with Visible World -- both will reap market share growth from their early moves. But it will be many years until we see that capability available at scale by all providers.
Not so when it comes to applying the unprecedented consumer data now available (massive amounts of set-top box viewing data versus small panel samples for ratings) to better scheduling of ads in the linear TV feed. Already, agencies are using tools like Kantar, TRA and Rentrak to find concentrations of target and measuring audiences according to day, time, network, program and geography - and at levels of accuracy 1000X that are available from existing ratings services.
Most of the value of 1-to-1 targeting on TV can be achieved by segmenting and delivering "mass customized" TV spots -- rather than profiling viewers for the perfect spot. Delivering ads to broad, anonymous segmentations of viewers likely to be "animated film fans" or "Friday night revelers" represents an enormous improvement over how TV ads are delivered today. Whether it represents 80% of what 1-to-1 can deliver, or 70%, or 90%, I don't know; but it's a lot. This is critical, because the cost side of the equation -- the systems, operations, additional creatives, measurements, etc. to deliver 1-to-1 -- is more than 5X the cost, as compared to scheduling mass customized ads in the linear TV ad feed today.
Virtually no privacy implications
Most importantly, the use of anonymous, aggregated consumer data to create broad anonymous market segmentations that can be used to better schedule and buy linear ads by time, day, show, network and geography creates no new privacy protection issues. It's no different than a new form of consumer demographics or psychographics. It involves no profiling. It involves to personal data. It doesn't require the ability to relate data to particular people or devices.
Do I think that the industry should not develop and invest in 1-to-1 ad systems and techniques for TV? Absolutely not. I am a big fan of the investments that folks like Cablevision have made, and the work that folks at Canoe Ventures, WPP/Group M, StarcomMediaVest, Direct TV, Dish and others are making. Some day, these systems and this work will reshape TV advertising as we know it. However, in parallel, it is important for the industry to find very privacy-safe ways to crawl before it runs. That is why I like leveraging broad, anonymous segmentation for everything it can deliver. What do you think?