By the end of 2016, IBM plans to buy the entirety of its programmatic advertising using Watson.
The company began experimenting with Watson to buy some of its advertising programmatically about a year ago, and the results were stunning.
According to IBM, using Watson to buy digital advertising lowered the average cost per click by 35%, in some cases reducing it by up to 71%. For a company that spent $53 million on advertising in 2014, the cost reductions will have a deep impact on the bottom line.
As IBM continues to use Watson for digital ad buys, the better the computer will become at doing so. As IBM’s vice president of marketing and analytics has stated, “What makes this really exciting is the system learns. That’s the essence of cognitive.”
In another use of Watson, The Weather Company, which is owned by IBM, announced in June that it plans to release a novel ad unit this fall enabling consumers to ask Watson questions "via voice or text and then receive information about the product or offering.”
However, some caution that we need to be wary of IBM’s genius data machine. A New Yorker piece from November 2015, "The Doomsday Invention," contemplated what AI will or could do to civilization. According to a senior IBM executive, “The separation between human and machine is going to blur in a very fundamental way.”
But, contained to the advertising industry, the use of Watson will enhance campaigns, lower costs, and inevitably free up cash to aid in developing varied aspects of a business beyond ad buys.
Too bad the metric is on clicks - which is a non modern way to understand an individual's engagement. When IB does figure it out - nd they probably will - Watson will be a very compelling way to buy media.
Very interesting. I've been thinking about how to do this for a bit now (programmatic ads, automated bidding, etc. via Watson), and it's not as straightforward as you'd think. And to what extent is it incremental to other methods? IBM doesn't share how it was doing ad purchases before ... it's easy to get big increases in efficiency if you start from a very basic level. Beware of over-hype on the efficiency gains ...
Even if some automated buying system that is mainly CPM- or cost-per-click-driven makes a huge "improvement" in targeting efficiency, though often at the expense of other factors like- content quality, audience engagement, overall reach, etc----it is unlikely that such gains will continue from one buying cycle to the next. Most likely you will get a big CPM improvement the first time but garner much smaller CPM reductions over time---at which point the penalties for not factoring in other important variables becomes more evident---and painful. Seemingly cost efficient ad campaigns do wear out. Really.