Marketing

Solving the intent data problem with journey management

Mark Smith, president at Kitewheel, says retailers are swimming in customer data for creating fairly detailed customer profiles. And, with the right analytics tools, they can extrapolate from this data to generate maps of buyer journeys and ideal customer profiles for more effective targeting.

Photo by istock.com

May 28, 2020 by Mark Smith — President, Kitewheel

Companies today are swimming in customer data — data on their preferences, online behavior, buying patterns, and much more. While sourcing all relevant data into one place can present brands with real challenges, in principle, every company has enough data at their disposal to create fairly detailed customer profiles. What's more, with the right analytics tools, you can even extrapolate from this data to generate maps of buyer journeys and ideal customer profiles for more effective targeting.

However, as valuable as this historical data can be, the most important customer action isn't something they did in the past. It's what they are going to do next. Companies strive to predict these next steps based on behaviors that resemble models of customer behaviors (customers who did x and y often do z) or by trying to infer intent from specific behaviors (customers who visit product review websites may intend to make a purchase in the near future). But how effective are these models at predicting what next steps a customer might take?

Some claim it can be very effective. And a raft of vendors have sprung up offering, at least in their view, actionable insight into your customers' next moves. When you take a closer look, of course, what these vendors offer is little more than inferential clues regarding where a customer might be in their journey. We see this particularly in the B2B world where review sites and sites specializing in content syndication essentially sell visitor and download data as signals of intent.  

To be fair, this may be adequate in a B2B scenario. Buying cycles in B2B are more drawn out and research is a critical part of the buying process. If you see someone doing research, it's not too great a leap to infer that they are “in market.” Individual consumers, on the other hand, don't have procurement standards or approval chains slowing down their purchasing decisions. The purchases they make can involve research, to be sure, but they are more likely to be impulsive or, at best, habitual. Without an in-depth understanding of customer journeys or a comprehensive framework to measure them, defining intent can be a formidable task.

Does that mean that brands in the B2C space should surrender all hope of predicting buyer intent? Far from it. What it does mean, however, is that to get close to buyer intent, you will need to be thoughtful about the way you work with your data, your data frameworks, and your models of customer behavior.

Building your own intent data

Organizations that implement journey management have a leg up on the intent data front. Through journey measurement and journey analytics, you can identify patterns that predict — and even influence — customer behavior. While there will always be some amount of inference involved when predicting the behavior of any individual customer, the more customer data you have, and the more confident you become in your customer journey maps, the less such inference will feel like guessing. It will begin to feel like prediction.

To get to the right level of detail in your customer journey maps, a measurement framework that relies on journey steps as its building blocks is the place to start. Aside from filling an unmet need in quantifying the effectiveness of customer journey initiatives, this approach to journey measurement provides you with a granular enough view to connect the dots between different parts of your customer journeys and, ultimately, to identify meaningful patterns and leverageable correlations.

Say, for example, an outsized number of customers who participated in a social promotion go on to visit your product pages. Of that cohort, a strong contingent goes on to make a purchase. By treating each of those interactions–the promotion, the website, the purchase–as a journey step in your measurement framework, you can begin to recognize the journey flow and, as a result, it becomes possible to influence it. More importantly, by identifying the journey steps that are most likely to lead to a purchase, you begin to compile your own storehouse of intent data.

Making it real-time

There's an inherent risk, however, in relying on this type of intent data, based on modeling journey steps: there are many variables that can influence customer behavior. Even the most loyal customer who's made repeat purchases can, from all appearances, seem to be on a positive trajectory, only to have a bad experience sour the relationship.

It's critical to remember that, even if your models are highly predictive, the customer is not following the model, they are following their own needs and wishes. It's also the case that, unfortunately, there can be variations in their experience with your brand. For this reason, to predict behavior or infer intent, you need to be able to monitor as many brand touchpoints as possible and, ideally, respond in real-time if something goes awry.

Let's take, for example, the extremely loyal customer mentioned above. Say they are a user of your subscription-based DTC retail service. They've been with you for years, have made several purchases, and seem ripe for an upsell to an upgraded level of service. If they have a negative experience, then look at the cancellation terms on your website, and then immediately call customer service, you could be in for trouble. For example, if customer service lacks adequate insight into the customers past history, or underestimates the urgency of the matter because they were unaware that the customer had already visited the cancellation page on your website, they could make a misstep resulting in irreparable harm to this customer's loyalty and the loyalty of anyone they might influence.

A comprehensive approach

While predicting behavior is valuable, it is even more valuable to be able to influence it in real-time. That requires having customer journeys accurately mapped, a framework to support journey measurement, and real-time decisioning capabilities. One without the others leaves too much to chance. As we said, any model is just a model. They can help you predict the future, but as soon as a customer deviates from the model, you need to be able to respond in a way that gets them back on track.

Finally, customer behavior patterns change over time. By focusing on comprehensive data collection and continuous analytics, you can not only respond to customers in a timely fashion, but also evolve your models so that they keep pace with caprices of customer behavior overall.

Mark Smith is president of Kitewheel.