Technically, the segmentation was created using a K-Means algorithm. K-Means is a simple unsupervised machine-learning algorithm that is able to unravel hidden patterns, correlations and even behavioral tendencies autonomously within data.
In short, the algorithm starts by plotting each guest’s answers in a 15-dimensional space, one for each question. Then, k-centroids are plotted randomly. From there, the following two steps are repeated in an iterative process:
- Each guest is assigned to the closest centroid. The distance between a data point and the centroid is most often evaluated using the Euclidean distance (‘straight-line distance’).
- The mean values of all the points belonging to a certain centroid are calculated, these become that centroid’s new location
The algorithm keeps on repeating these steps until there is no change in the centroid values or a previously determined maximum number of iterations has been met.
Upon completion, the algorithm had minimized the within-cluster variance: meaning that each group is as homogeneous as possible, sharing similar characteristics and behavioral tendencies. Based on this information, we were able to stick a label on each cluster: one of our five music vibes.
Whether or not our little educated guess about a guest’s musical preferences was spot on or not, we probably will never know. What we do know is that all vibe-members do share similar sentiments towards the good old 90s. Grunge-members will have a shared love for Super Mario, as well as a shared hatred for dancing the Macarena. Hip-Hoppers will vow by Napster, while Brit Poppers in general prefer to support an artist by paying for their music records.
A simple exercise like this already uncovers what segmentation is about and what it can do for your business. Segmentation allows a marketeer to better understand his audience and to better target each individual according to his personal preferences and needs. As an online reseller, you do not want to highlight Kendrick Lamar’s brand new CD to those free-riding Hip-Hoppers. You might however, want to highlight that pair of Dr.Martens when a known Brit Popper is visiting your website.
Applied to a marketing context, such information can be extremely valuable. The 80/20-rule states that 80% of sales are due to only 20% of customers. Knowing which customer profile is most valuable to your business and knowing how to target them, is key to a successful marketing campaign. If a web visitor has similar characteristics, interests and/or behavioral tendencies as your top customers, it is highly likely that he or she may become a top customer as well.
Moreover, segmentation facilitates personalized marketing. While personalized content is not feasible in most business contexts, targeting each customer segment differently may offer a near-personal experience. For example, different emails can be sent to different customer or prospect profiles.
Customer segmentation offers numerous possibilities to take your marketing to the next level and in many cases, it can be realized with minimal effort. It really is a great opportunity to get to know your customers and target audience better.
We - at The Reference - encourage you to define your business case. Feel free to propose it to us. Chances are that our data team might just offer a valuable solution.
I HAVE A PROPOSITION YOU CAN'T REFUSE
Do you want to stay informed and get access to more interesting, data-related services such as our data-to-value framework? An inspiring set of tools that help guide your company from the ideation to the operationalization of your most valuable data opportunities. Then please, do subscribe to the newsletter. You can find the subscription form underneath this post.
Thank you for reading.