March 8, 2017
“Here comes our robot driver!” my friend excitedly shouted as our Uber pulled up. On a recent trip home to Pittsburgh, we had been lucky enough to get picked up by one of Uber’s self-driving cars. Of course, there wasn’t an actual robot in the driver’s seat, but this didn’t stop us from dreaming of Mr. Roboto chauffeuring us home. Instead, Uber engineers greeted us from the front seats, and informed us that the car would be driving itself. This taste of the future was all thanks to machine learning.
Machine learning is a way to deal with situations that humans are not well-equipped to handle. Although humans can drive well, we aren’t as good as a computer. Google’s self-driving cars now have the equivalent of “300 years of human driving experience” and have only been found to be “partially” at fault for 1 accident. Humans simply can’t compete with a machine's capacity to ingest, analyze, and act on large volumes of data.
Monetate’s Intelligent Personalization Engine is utilizing machine learning to help deliver individually relevant content to shoppers. Most businesses have a wealth of data about their customers and their preferences, but pulling this together to deliver an engaging message is difficult. At best, marketers can analyze historical data to display messages to segments of shoppers. The problem with this approach is that it uses older information and doesn’t see through the segment to the diversity that can exist within those segments. Monetate’s Engine, however, can observe trends across all of your data points to deliver relevant content in real-time to shoppers on a 1:1 basis. No marketing or IT department could reasonably operate at this scale or speed without machine learning.
Trusting a machine with your business is a tough thing to do, however. In the case of a self-driving car, you’re literally trusting your life to the machine’s algorithms! How can you be sure the machine is doing what it says it will do? Why should you trust it? Uber makes its case by putting a screen in the back seat that displays a live feed of exactly how it is interpreting all of the current data. The screen allows you to see a real-time computerized model of the street you’re driving on, so that you can feel more at ease that what you are seeing is the same as what the car is seeing (take a look at it here).
Similarly, Monetate’s Engine offers transparency and doesn’t work as a “black box”. At any time, you will be able to see how traffic is being distributed based on the most important observed data points. As your customers change, the models will change, but you will always be able to see how the Engine is making decisions in real-time. Because the decisions made in the moment don’t necessarily reflect how traffic was, or will be distributed, it’s important to note that this does not replace post-hoc analysis (You wouldn’t want to drive your car based on one snapshot of the road, right?), but it can help you make directional decisions.
For example, if you see that a high number of mobile users are responding well to more simplified content, perhaps any future mobile content you create can take this design aesthetic into consideration.
Machine learning has been a concept for decades, but is only now becoming more widespread across various industries. While waiting for our Uber, another friend in my group - who would not be joining us for the drive - noted with disbelief that he had yet to take part in one of these futuristic forays. As we drove away, he watched with envy, wishing that he was coming along.
With Monetate’s Engine, machine learning is poised to help businesses take the next big step in 1-to-1 personalization - don’t get left behind!