As a company of marketers who love a healthy dose of storytelling, we can’t help but interview interesting people about what they’re most passionate. We snagged IBM Watson Marketing Lead Brian Rants and chatted about the way marketing automation is changing, where it’s already been, and what nerdy thing we should be looking forward to.
Look Listen: What is your history with marketing automation?
Brian Rants: I started out building websites in coffeeshops in Denver, and my personal movement into marketing automation was an evolution that seems to mirror the industry. Initially, my agency grew from building and marketing web properties (SEO, SEM).
But as businesses shifted more of their business development and customer conversion towards digital, it became essential to automate key programs. So, not just spending on the initial engagement, but nurturing the customer during consideration, helping them with those gentle pushes toward conversion, then delighting them post-engagement to turn them into loyalists. It’s easy to say this now, but at the time it looked more like businesses starting a monthly e-newsletter. Then a year or two later, they might want to add an email drip program to previous purchasers in B2C or to cold prospects in B2B.
Much of my growth and experimentation with tactics actually occurred at a Denver direct marketing firm where I was brought in to create their email marketing automation division. While I could share from my background in digital, I learned just as much from experienced direct marketers for whom creating and optimizing drip campaigns, calls-to-action and response rates was normal course. For me, it was this melding of channels, maturing of digital marketing, and the combination of left and right brain that gave life to marketing automation.
What is AI’s role in marketing automation? What new outcomes are now possible?
AI is helping marketers make better decisions, faster, at greater scale, and with unparalleled accuracy. People think of AI like a robot, but the way I see how it’s applied in Watson Marketing, it’s more like Captain America’s suit. It heightens your senses, makes you orders of magnitude faster and more powerful. Also, like Jarvis, Watson interacts with you in natural language.
Rote, yet essential, tasks like tagging images could take an entire week of a resource’s time over the course of a year — or more likely, they’d just never get tagged (and you end up with names in your system like “496234-LARGE-final-FINAL-v2.jpg”). Watson leverages visual recognition to understand what’s in the image, and recommend tags.
Another example: Your boss just asked you to pull reports for an all-hands meeting about your up-sell campaign this year vs. last year, which for many marketers is a multi-day process full of manual adjustments. With Watson, just ask and you’ll have a report in seconds.
This lets you shift your time from foundational tasks into higher-order optimization. What’s the relationship between sent time and open rate? How are channel, device type, and email type (triggered vs automated) affecting performance? AI is helping marketers become more effective versions of themselves.
What’s new for MA outside of AI?
The previous question is where we get to have fun, and talk about the Avengers. In this question, I think what’s new is what’s old — integration and execution. Without integrating your data sources, how will you effectively coordinate execution? If you can’t understand who your customer is across channels, how can you meaningfully understand what the performance is?
Watson Marketing built Universal Behavior Exchange to federate identities, to exchange behaviors and audiences, across paid, owned, and earned channels. Other providers have also emerged in an “infrastructure as a service” role. Whatever you use, the problems in your data layer will only become exacerbated with layers of automation — so start with a great agency like Look Listen to create an effective data architecture and build out advanced automations! Just as importantly, work with a partner like Look Listen to build (uh-oh the “p” word) processes for deciding what do with the results. We have beautiful multi-dimensional reporting in Performance Insights, but what’s your internal process for efficiently making mid-course adjustments based on performance feedback?
What’s old is new: As you automate more and more programs over greater data, integration and good old-fashioned execution practices become essential.
What are you nerding out about in marketing automation?
Data science without the data scientists! If you had a large company with a mature marketing analytics organization, you could always build predictive algorithms to detect churn, determine who was ripe for an offer, and so forth. With Watson Marketing Insights, we’re leveraging Watson to look at CLTV, churn and engagement — and predict movement before it happens. That means putting predictive analytics in the hands of email and mobile channel marketers, which seems to me to be something like Iron Man’s boot jets…