Some time ago the folks over at the Analytical Engine had a post about Data Driven Persona Development. I loved the way they described taking a very qualitative approach – persona development – and adding some analytic rigor to it. Given my interest in using analytics for segmentation and in developing different websites/experiences for different segments, I thought I would link to it and expand on it a little.
So, let’s assume you are developing a website or some other interaction for your customers. Using descriptive analytic or data mining techniques you can develop material, statistically significant groupings or segments. Having identified the segments that matter to you, you can develop personas for these segments as described in the post. When you start developing the interaction, you can use the personas to drive design decisions, typically resulting in some different interface or interaction choices for each. Now, because you know the segmentation model behind these personas, you can also generate the rules you need to put someone into these segments and thus direct them to the design that matches their persona. This decision – what’s the most useful persona to use for this customer – is thus automated and driven by your segmentation model. Lovely.