April 5, 2017
The great fantasy of marketing is that, in an ideal world, we would be able to produce unique creative for every single customer. We would know them as individual people—their preferences, histories, and idiosyncrasies—and we could help them connect with our brand by speaking directly to their perspective, using the story that is most relevant to them.
When I first started working in this space in the mid '90s, a lot of people thought of that as the reason to be excited about digital: it was a place where we might eventually be able to create a unique experience for each customer. That’s how how our industry still defines personalization. The allure of completely individualized creative has shaped our methods for decades.
Since nobody has the means to build millions of unique experiences, we've had to to seek alternative solutions to better reach our audiences. Testing and optimization has worked very well for lots of marketers, including our clients: this method allows us to learn about an audience and use our findings to improve our offerings. First we hypothesize, we generate a couple of alternative pieces of creative to test out, and then we measure which designs perform the best on average for the whole audience. The winning design is then served to everyone who visits the site. However, when you’re only offering one option to everyone, the benefits of a/b testing become increasingly hard to scale as many ideas work well for some customers but not for others..
This is where we naturally introduce segmentation, which is the first step of moving from homogenous to heterogeneous design: instead of offering the same thing to every customer, we can use a few defining characteristics to split our audience into two or more groups and design specifically for them. This allows us to optimize separately, to great effect. Our clients have seen significant gains from this approach, and they’ve also done a better job for more individual customers in their audience. However, just like with testing, there’s a point of diminishing returns: after the audience has been divided enough times, ROI drops at the point when producing creative for each segment becomes too resource-intensive compared to the potential returns from a too-small segment. Eventually, segmentation reaches the point where it’s straining to approximate that fantasy of individualized creative for each customer.
Both testing and segmentation are still the most appropriate choice for certain use cases in many businesses, depending on their specific objectives and what resources, limitations, and customer bases they are working with. However, neither one scales to, or even approximates, the ROI of 1-to-1. When you reach the point where your optimization method hits the point of diminishing returns, it’s time to add a new approach.
It’s time to redefine how we think about personalization. As an industry, we’ve been limited by our fixation on the artificial goal of creating a “segment of one.” At Monetate, we believe that the real power of personalization lies in a different approach to the problem. Individual creative is obviously unachievable, and that makes it an ineffective goal. We’ve realized that the dream of 1-to-1 is very much achievable when we focus instead on making unique, individual decisions for each customer. To hear more about how to achieve 1-to-1 and the ROI that comes along with it, read our whitepaper, Demystifying Personalization.