As warmer weather approaches, we have cold, tasty ice cream treats on our mind. It turns out, we’re more than daydreaming. A great analogy for the benefits of 1-to-1 ecommerce personalization lives in a simple example of ice cream flavors. We can illustrate how different strategies stack up by imagining we’re giving ice cream to a stadium of fans. Who doesn’t love a good ice cream analogy?
Picture yourself at a baseball game celebrating fan appreciation day. Our goal is to provide a tasty ice cream treat and make everyone in attendance feel special. In an a/b testing approach, our creative team would generate a series of alternative options and then figure out which one would perform the best with the audience. Maybe they make vanilla, chocolate, and strawberry ice cream to test as possible flavors to serve to the whole crowd. We would poll the audience to see which one was preferred by the majority. In this scenario, perhaps we would discover that 45% of people wanted vanilla, 40% of people wanted chocolate, and only 15% of people liked strawberry best. Vanilla becomes the winner of the test. And though less than half of the total audience wanted vanilla, it still got more votes than the other flavors—so it becomes the only offering. Unfortunately, this outcome leaves 55% of people unsatisfied—more than half!—even though vanilla was the most popular option of the three.
Obviously that isn’t an ideal solution. While dynamic testing technology can adjust to promote the alternative that is performing best, if you are only serving one offering to the whole audience you will be inherently limited by a homogenous approach: you will lose anybody who doesn’t fit into the average.
We could capture a lot of those other 45% by identifying segments: that would allow us to split the audience into chunks and target them separately, assigning a different flavor to each group based on profiles of the members and their preferences. We could use data about people’s past ice cream consumption, or a correlation between hometown and flavor preference, to guess what kind of ice cream they will want today: all members of the first group would get vanilla, the second would get chocolate, and the third would get strawberry. Even if we were wrong about some individuals it would be a better experience for most people.
While an improvement, this is not the best we can do. Someone who normally goes for chocolate might be in a vanilla mood, or there might be data that we didn’t have that would change the implications of that person’s past preferences. Even though we’d be better off than we were producing one average flavor for everyone, we would still just be designing for increasingly granular averages. Eventually the segments would be so specialized that we’d be designing a custom flavor for almost every person, which is not sustainable.
These methods can deliver excellent results for businesses that need a boost in the short term, but they don’t truly scale to the individual level, or “1-to-1.”
That’s why Monetate’s new product, the Monetate Intelligent Personalization Engine, is exciting: instead of serving a single option and losing those outside the average, or breaking our audience into labor intensive micro-segments, the Engine decides which of the available options is best for each individual customer based on their data. You can maximize your return on creative by making sure that you serve the best choice to the each person—every single time. We’ve redefined personalization: it isn’t about producing infinite creative options, it’s about being smarter about how you decide on the right creative for each individual.
With Monetate’s Engine you can decide how many flavors you want to produce—maybe it’s even just the original three—and then serve the right one to the right person every single time. To learn more about how the Engine is changing the definition of 1-to-1, read our white paper, Demystifying Personalization.