November 1st, 2017
Every digital marketer that’s paying attention is either wary, concerned, or deathly afraid of machine learning and artificial intelligence.
These systems feature absurd processing power and instant analytical capabilities. They eat big data and crap hyper-targeted marketing. They take no breaks or vacation days, and spend no time screwing around on Facebook (except to ingest behavioral insights to make themselves smarter).
Various tech behemoths have branded their marketing AI in their own image. IBM (#client) has Watson, of course. Salesforce (#client) has Einstein.
I’ve always thought Watson and Einstein existed to make these very complex systems more understandable. “Oh, that intersecting nexus of dozens of data feeds? That’s just Watson!” But now, I wonder if part of the benefit of this anthropomorphic approach to explaining AI is, perhaps unintentionally, to prevent marketers from totally freaking out about the future. “Watson isn’t going to take my job! He’s so funny and charming!”
This is the marketing technology equivalent of distracting a child with a doll just before drawing blood.
In truth, I do not believe AI (IBM calls it “cognitive”) is going to wipe the digital marketer off the map like a herd of grass-munching brontosauri. As I said at the annual Sitecore (#client) user conference last week, when the robots do all the hard, tedious, time-consuming digital marketing work, the strategist becomes king.
Because here’s the thing: For the foreseeable future, these systems must be TOLD what to do. Yes, machine learning, by definition, dictates that they get smarter and better over time. But they first must be taught what to look for and where—by us.
If your job entails a lot of manual testing, list splitting, tweaking, optimization, and scorekeeping, you must learn how to manage the AI robots (and soon), or you may end up superfluous—as vestigial as our own appendix.
The Social Media Robots Are Here
For the last two years or so, most of the work I’ve done and been exposed to in the AI-powered digital marketing world has been focused in these areas:
- Email testing and optimization
- Web content testing and optimization
- Mobile testing and optimization
- Behavioral data pattern recognition and ID of “best” and disenfranchised customers, as well as likely advocates and influencers
To date, social media (other than social listening, of course) has been largely immune to the effects of machine learning.
Yes, you can A/B test organically on Facebook. Social ads allow for real-time optimization. But the kind of serious, one-to-one interactions that marketing technology companies have spun up for email, web, and mobile haven’t really been doable in social. After all, it’s social—it’s right in the name. It’s more of an art than a science, right? You have to nail the right tone, tenor, and timing.
Nope. Not anymore.
Persado is a fascinating company I’ve been monitoring closely since I did an early demo with them five years ago. They use a dizzying array of data inputs and DMP relationships to create highly targeted, massively specific content pieces inside email, on a home page or landing page, etc.
Their core assignment has been helping e-commerce companies boost their conversion rate by dramatically boosting relevance and resonance of digital copy and imagery.
But last week, they launched Persado One which allows digital marketers to customize messaging and look and feel at the individual level, rather than just the customer segment level. Even more importantly, Persado can now deploy those one-to-one campaigns as social media ads and—get this—craft them based on historical emotional engagement.
My head is swimming with use cases for this technology. Imagine the Grand Ole Opry (#client) wants to use Facebook ads to sell more tickets for upcoming performances. Based on users’ historical engagement with messages of different tones and tenors, Persado can combine copy and images and Frankenstein together the perfect ad for EACH USER.
If, based on your website, email, and social media behaviors, the database believes you like to go to the Grand Ole Opry with your family, you’ll see a Facebook ad emphasizing that benefit. However, if your history indicates you are more into seeing classic Bluegrass artists live, you’ll see an ad with those attributes highlighted.
This is either the beginning of the end, of the end of the beginning. Either way, the days of social media success being about human-powered emotional intelligence are waning.
All hail our robot masters.