Turning personas into data-driven profiles for segmentation

Personas have been a helpful tool for supporting user experience design for some time, allowing brands to identify the needs, goals and problems of an audience segment in order to better meet their wants and needs along the customer journey.

Often though, these fictional characters meant to represent different segments of users within a customer base are simply too abstract, lacking the rich information necessary for teams to truly adopt or drive any real meaning from them. Left untouched, personas become a fading reference point within the org, never fully influencing the decision-making process.

In order to create real value, personas must go beyond mere generalizations, stitching together different data sets for a more grounded understanding of each customer group. Only then can they be effectively used for aligning on and powering the delivery of truly personalized experiences.

Largely derived from preconceived notions of who we think our customers might be based on our own personal contexts, typically, personas end up looking a bit like this…

While it might look nice and attempt to humanize outreach, these oversimplifications provide very little insight into what kind of experience a business should deliver to these audience types aside from some very general demographic information.

What actually makes users different from one another has a lot more to do with just age, gender and geo, incorporating device specific information, motivations and interests.

Let’s get a little more granular with Daniel, diving into his background, potential preferences, time on site and what the most ideal experience for him to receive might be:

Consider this the hypothesis portion of your persona building. Not concrete truth, it’s an exercise in fleshing out specific audience theories for validation later on while analyzing the impact of an audience on conversions and revenue.

Even with the context above, it’s still too difficult to create an audience for the purpose of delivering unique experiences. While it forces a brand to put itself in the user’s shoes, it lacks truth and is grounded entirely upon imagination. This is when data insights must come into play, defining the persona and his or her actual motivations.

Using the following scale, information can be tapped at varying degrees of accuracy to better serve the needs and wants of your customers — the more sources of data insights, the better.

The options to create accurate and interesting audiences are endless, and any combination between sources can lead to a meaningful segment.

Here’s an official breakdown of each source ranked by accuracy and size:

Ideally, a balanced combination of each source of data will produce maximum relevance in proportion to its size and accuracy.

From there, a persona’s character can be truly defined for use as a segment. Based on Daniel’s original description, we can confirm the following information about him:

With a complete picture, each persona created becomes the center of an experiences’ design, allowing brands to more effectively navigate what touchpoints to influence, when and how. Based on real information and not just pure assumption, these data-driven personas or macro segments provide a stable jumping off point for potential experiences along a personalization testing roadmap.

Upon tailoring experiences for these larger subgroups of visitors, smaller audience groups will emerge, providing additional opportunities for experience creation, optimization and revenue generation. And while moving from a macro approach based on the persona work above to micro segmentation requires a lot more time, effort and resources, the output of doing so is even higher-yielding results.

By analyzing trends within these data insights, brands can identify and define smaller, finely tuned customer segments who are contributing the most value, a huge step towards achieving revenue growth. Now, with a great deal of information about their buyer’s preferences, context and intent, serving these users becomes much easier. The hard part is then customizing messages and content for each small segment and thinking carefully about how to optimize the user experience and flow being created.

However, with the right tool, this process can be made simpler. Using predictive targeting, brands can lean on machine learning to analyze the performance of an experience and it’s collective variations or treatments across different audience segments in real time to identify the best possible one for any audience, large or small. Able to deploy with confidence, brands can begin executing on a truly individualized approach — and it all started with a few basic personas. 

This post was co-written by David Ochman.