Even seemingly simple journeys are complex, multidimensional constructs. It might help to compare journeys to something like DNA—not the static diagram of a middle school textbook, but a 3D model of intricate protein pairings that are rotating, spinning and changing in unknown ways over time. One can grab a bit here and there to get at readily available customer journey metrics, but the simpler that measurement, the less it describes the entirety of the object.
Journey analytics can provide meaningful and productive insights from relatively simple measures. But as anyone versed in machine learning will tell you, the richer and more complex the data you feed in, the better the insights coming out will be. Our intent is to use journey analytics to predict human behavior—to algorithmically replicate what is going in a customer’s head to enable your business to make split-second decisions.
Every time a person makes a choice—vanilla or chocolate—their simple answer is the product of remarkably complex data, input that spans time, place, and varied experiences. Choosing chocolate at age three is not the same as choosing chocolate at age thirty. Using very complex input on what appears to be a reasonably straightforward question will make the answers to that simple question much better.
The closer you get to modeling human behavior, the better the predictions should be, but such modeling requires tools that can make complex input manageable. Organizations themselves mature into users of complex customer journey metrics. That is important because this developmental process helps an organization’s data science and executive teams get the most from their journey analytics systems over the long haul.
Start with High-Level, Simple Measures
The common denominator we find among organizations that strive for that level of sophistication is an evolutionary path in developing their journey measures. It starts with high-level, simple measures fed into machine learning, as well as into dashboards and presentations for executive decision makers. The first insight might be as simple as “I see 50% of our customers are on our website at least three times per month.” The inevitable “why” requires digging deeper and using more advanced measures. Each subsequent question leads you further into complex customer journey metrics, whether the questioner is an executive or a machine learning system.
Two shifts occur as an organization progresses from simple measures through intermediate and advanced measures. One is a change in the level of granularity. At the elementary level, you are observing from a high level so that all that’s known is the fact that “a customer visited our website.” At the intermediate stage, the tendency is to focus so tightly that you get both increased useful measures and more noise. With increased sophistication, you can begin to balance the level of detail and value of the measure, settling at a level of optimal value.
Applying Time: Elementary to Advanced
The other shift is the introduction of time to the measures and incorporating increasingly sophisticated applications of time. It’s easy to take a snapshot in time, such as finding out how many people were on the website in March. A slightly more sophisticated measure would be to look at the trend from January through March. The change in mindset is to move beyond static snapshots toward a continuum that recognizes that change happens over time. It is a mistake to consider classic customer segmentation attributes as static. Who are you, where do you live, how much money do you make, what do you have, which of my products do you own, what competing products do you own? The answer to each of those questions changes over time. At advanced stages, you intentionally consider time as the customer journey metrics itself: days between balance inquiries, or length of personal interaction chains. Advanced time measures will provide for better understanding and behavior predictions.
Walk Before You Run
Every organization is unique and will have its own organic process of change. Even so, this high-level evolutionary path is important. Business teams quickly and intuitively latch onto advanced journey analytics use cases, like improving customer retention. However, organizations find the best return when they restrain themselves from jumping too far ahead. Start with simpler measures, show value quickly and consistently, and build support and capability for more sophisticated measures as the whole organization iterates. Managing a long-term strategic program will ensure clear and actionable results at every step of the way.