This article is part of Faraday's Out of the Lab series, which highlights initiatives our Data Science team undertakes and challenges they solve.
When it comes to understanding how prospects and customers engage with your brand, journey mapping is an essential tool. This solution identifies useful patterns in customer behavior and helps brands further optimize the customer experience.
Faraday has recently fielded many questions from clients around how to make more accurate product recommendations and create a more personalized website experience for users, both of which tie directly into journey mapping. And the number of use cases only continues to rise. So how does journey mapping work?
What is journey mapping?
Many brands think of the customer journey as a series of interactions between a prospect and a company, from the first time the prospect becomes aware of the company until the moment they complete a purchase. But equally important for brands is understanding how the customer journey continues after that first purchase. It costs much more to acquire a customer than to retain one, after all. That continuing customer journey is what we want to map here.
To map out customer journeys, we look at the networks of relationships between products and events. These relationships manifest when one product is bought after another or events consistently happen sequentially. To model and visualize these relationships, we use directed networks, which are connections of nodes and directed edges.
Let’s break that down.
Twitter’s social network has a similar structure, so it makes for a good example. Nodes represent Twitter users, and directed edges represent the action of users following other users (i.e. the relationship). Translated to a business-to-consumer situation, nodes might represent products, events, or actions, and directed edges are the chronological sequential relationships.
Let’s take these networks and go macro and micro!
Journey mapping from a macro perspective
Network communities are groups of densely interconnected nodes. For example, within the Facebook social network, a group of friends shares densely interconnected edges with each other and sparser connections with users not within the group. Our aim is to extract the macro community structure of the network using something called a nested stochastic block model (NSBM).
Using an NSBM, we can observe broader trends about community membership, with each community containing chronologically proximal products, events, or actions. For example, in the orange community, if we find several water bottle products and a trending eco-friendly straw, our client can leverage this relationship to inform their marketing efforts. The community perspective provides a macro view of the relationships between many products.
Journey mapping from a micro perspective
In addition to understanding broad relationships between products and events, we can zoom in to gain a more granular understanding of specific node interactions – the sequence of purchases or events. To do this, we follow the most informative edges out of nodes of interest. But how do we decide what is actually informative?
Let’s say we want to know what a customer may purchase next after a certain transaction (we call this sequence a transition). We would measure the probability of that purchase happening, along with its lift, which adjusts for the popularity of the subsequent product, making the transition along with the lift of the transition, which adjusts for the popularity of the subsequent node.
For example, if transitioning from an eco-friendly straw (Product A) to a water bottle (Product B) has the highest probability out of all transitions from Product A, but Product B is the store’s most popular product, then you would expect this probability to be relatively higher and would doubt the utility of the measurement. This isn’t very useful information because, despite the fact that Product B is the most popular, that doesn’t mean the customer will buy it. They probably already know about it, so we’ll look to less probable products they may turn to for their next purchase.
There may be a product C that, although less popular, practically always follows product A. Considering both probability and lift of transitions contributes to our understanding of what is informative.
How to enhance the customer experience
Journey mapping helps brands understand the sequence of actions a customer is likely to take – and it has strategic implications. With this information in hand, you can optimize your customer experience by delivering product recommendations and promotions to customers most likely to make those purchases.
This type of outreach not only smooths out the customer journey, but can also increase the trust customers have in your brand – and thereby increase their lifetime value. If you're serving up offers and promotions for products they're already considering, or products they hadn't heard of yet but are delighted to discover, they'll be more likely to keep engaging with your brand over time. In that sense, journey mapping is a powerful strategy for growing customer loyalty.
Want to know more about how you can track customer journeys and lifetime value? Check out Faraday's Maximized Customer Value solution.