Faraday personas are developed using your first-party customer data, our third-party consumer data, and unsupervised machine learning. At a high level, the machine learning algorithm sorts your enriched customer data into distinct groups, which are ultimately used to define your personas. Here's how it works:
The effectiveness of any machine learning application is heavily dependent on the training data. When it comes to Faraday's applications (descriptive and predictive analysis of consumer data), transactional data isn't enough to truly describe who your customers are as real people, let alone predict behaviors.
That's where third-party data plays a crucial role in any analysis and why data enrichment is one of the first steps in our persona development process.
Data clustering is a descriptive analysis technique that uses unsupervised classification algorithms to sort data into distinct groups based on any number of dimensions.
When clustering for personas, Faraday applies a version of the k-means clustering algorithm to sort your enriched customer data into groups based on a variety of consumer attributes from the Faraday Identity Graph.
The example above is easy to visualize. Now, imagine adding another six or seven dimensions (i.e. age, marital status, preferred shopping style, etc) to the graph. The added dimensions let us define the resulting clusters in a more meaningful way.
Persona definition is part art and part science. Considering where the clusters are centered against various dimensions helps us understand what makes each persona unique compared to the rest. It's a good place to start, but it's not the end of the road.
When we look beyond the clustering dimensions, we often find other prevalent attributes amongst certain clusters. For example, we might see that pet ownership is especially common amongst customers in cluster X, which should be considered when defining that persona.
Once defined, there's a lot you can do with Faraday personas. Check out the webinar below to learn more about personas and personalization with Faraday.