Breaking the Rules in Behavioral Targeting
In the constantly changing world of digital marketing, segmenting your audience using rules based on broad assumptions or unsubstantiated insights can work against your success. Unfortunately, many marketers rely only on limited rules-based targeting systems, without realizing that better options exist.
The reality is that given today’s complex relationships between businesses and their customers, traditional rules-based targeting systems are often not powerful enough on their own to give customers richer, more relevant experiences when moving through a digital universe.
We all know that even customers who might appear similar on the surface can ultimately have wildly different interests when it comes to what they buy and what prompts them to make a purchase. Instead of relying only on rules-based systems, marketers also need to tap into more automated, behavior-based targeting solutions built on powerful algorithms.
So what’s the difference? And why is at least knowing a little something about the underlying algorithms so important?
Adobe Target incorporates machine learning to give marketers a clearer, nuanced picture of site visitors. In real time, algorithms in Adobe Target look for as much information as possible about a visitor so the system can instantly display relevant, meaningful content. It takes into account a multitude of information to deliver just the right content, whether the system is considering an individual’s past session behavior, basic demographic information, participation in loyalty programs, specified interests, or other possible influencers.
A lot of digital marketing technology vendors offer rules-based targeting systems that display certain content to visitors based on assumptions. For instance, a website visitor from California might be greeted with an ad for sunscreen or shorts, while someone from a colder climate might only see jackets and other gear to keep them warm. Unfortunately, this approach treats large customer segments the same, shutting out potential opportunities. There’s no real intelligence behind the automation, even if vendors claim otherwise. To make rules-based targeting work better, marketers have to spend a lot of time defining rules and coming up with hypotheses—which is impossible when dealing with a virtually limitless number of customer backgrounds, interests, etc.
It’s important to note that rules-based targeting isn’t a bad thing—we employ the technology in many of our activities—but digital marketers should know it’s only one of the things they should be doing as part of their site optimization program. They need more automated modeling capabilities as well.
With automated targeting built on machine learning, marketers don’t need to manually manage and build models, and instead can spend their time on more valuable activities. Even the staunchest proponent of manual intervention and control would find it hard to argue the fact that for many marketing activities, machines are much better equipped to quickly move through thousands of variables and make correlations between things that—at least to us—could seem completely unrelated. Machine learning also has the advantage of getting better at learning over time; targeting accuracy improves as data is collected and analyzed.
We know empowering our customers to make smart decisions relies on our success at giving them smart tools built on powerful algorithms. To help us stay on the leading edge, our development teams regularly participate in rigorous academic modeling competitions, as well as continually refine our targeting solutions in our in-house labs.
A good dose of healthy competition with some of the world’s leading developers is a good thing. As is our focus on doing right by our customers—and your customers—which leads us to wanting to help you understand what and when rules-based versus more machine-based targeting is best. We’ll continue to explore this topic further in another upcoming blog.