The promises of AI are everywhere: security, medicine, education, finance and even agriculture. Getting comfortable with allowing lines of code to make decisions for us isn’t an entirely a new phenomenon but we are increasingly being conditioned to accept that the machine “knows best.”
It’s not always a comfortable position as a Marketer (or even as a human!), especially if you are a bit of a control-freak and must understand how “the sausage is made” to be on-board with any new initiative the big advertising platforms roll-out. Now though, is the time to get comfortable – to understand yes – but to be comfortable with that lack of control when and where it’s needed.
Trusting the Machine
Often we’d like to know what creative performed best during a campaign. After we review all creative against each other, noting which ad(s) had the best CTR, and which ad(s) had the most conversions, we can’t always say why ad “X” outperformed ad “Y.” The data just can’t tell us that. As humans we can make assumptions but at the end of the day, we really don’t know with 100% certainty.
People are different, people are fickle. What works for one person won’t necessarily work for another and what worked for that person today may not work on him or her tomorrow. We’ve learned to take the trends that we see and run with them, using them to inform future campaigns. Sometime we move in the right direction, sometimes we don’t. It is the goal of AI to move us in the right direction more often with less effort.
Over the past 5 years, Digital Marketing technology has grown significantly more adept at defining “audiences” and making those audiences available for targeting. But often due to fast turnaround times and/or a client’s limited access to creative assets for us to work with, the same creative is served to all of these carefully defined audiences. So we learn what works with one audience but not another. While this is a great lesson, it also means we may have missed the mark with the other audiences because resources were not available.
It is these gaps in Digital Marketing that make the promise of dynamic creative and AI ad delivery so tantalizing. Why worry so much about the “why” of what worked well in last year’s campaign when I can easily offer up not only an updated version of the winner’s elements but quite a few more different elements and let the system assemble the elements in the way that is best for that user at that particular instant. And why worry about what ad to serve to what audience when again, the system will do that for me.
Why indeed. As much as I’d like to believe AI has my back, it still needs help. None of us could ever take hundreds (thousands?) of data points and make a split second decision the way AI can but AI can’t do what we as humans do well – communicate.
The time we save letting the algorithms make the best decision in the moment should be time that we spend making sure we are providing AI with the best information to start with. This means worrying less about the demographics of our audience and more about how we are going to supply the machine with data points around the behavior of our target audience.
This means spending less time trying to mimic the perfection of print ads in the digital space and instead focusing on supplying a variety of messages both in the ads and on the landing experience that speak to the various ways people make decisions.
The Illusion of Control
See, the thing is we never had control. Customers have always held the cards; it’s just that now we get more bites at the apple. And that’s a good thing.
So get comfortable that you don’t have 100% control over how your ads might appear and please stop only defining your audiences by demographic information. People are individuals and the best thing AI algorithms can do for us is to get better each and every time in predicting what is most likely to appeal to that person in that moment. Learn to trust and to continuously test.
It’s a great time to be in Digital Marketing, even for those of us who are control freaks!