Customer Journey Analytics

While companies often view processes from their frame of reference, “cutting” processes up according to department, business objective, or other internal aspect, customers obviously do not act according to the same taxonomy and—from the perspective of the company—appear to jump from process to process, from department to department, and from channel to channel, making it difficult for businesses to truly follow a customer through his or her whole journey.

As such, many organizations have adopted the concept of “customer journey mapping,” or “customer journey analytics” in order to map out the customer’s experience and analyze it from initial contact with the company through the steps of engagement and finally into a long-term relationship. Such techniques are, for instance, frequently applied to analyze the “onboarding” of customers, where customers might first look up information on the web, rely on word-of-mouth, check out a vendor’s mobile app, make some calls, or visit branches before committing to a particular purchase of a product or service.

Making sense of the customer’s motivations and expectations can help a great deal in figuring out how to increase sales, uptake, and quality. Customer journey maps help to uncover gaps between devices (users move from their phone to their tablet to their computer), between departments (e.g., the customer gets forwarded from the sales to the invoicing to the support department, all working with different identifiers, which frustrates the customer) and between channels (e.g., a customer interacts with social channels such as Facebook, a website, mobile app, or call center and can jump between them). These gaps are often not apparent when thinking about processes from an internal, business-centered point of view, but stand out in a customer journey analysis, where the user or customer is put at the center of thinking.

More formally, a customer journey is a diagram representing the various steps the customers go through when engaging with a firm (Richardson, 2010). Figure 1 shows an example of a simplified customer journey for a mortgage sales process. It illustrates the various activities, states, and transactions that a customer can be in when buying a mortgage. Transition probabilities and time indicators are usually added for further enrichment of the analysis.

Customer journey analysis serves various business purposes. It can be used to get a clear and comprehensive picture of the overall process and highlight process deficiencies such as excessive processing times, deadlock situations, circular references, unwanted customer leakage, etc. It can also be used to verify if the process is compliant with both internal and external regulation.

From an analytical perspective, sequence rules can be an initial approach to discover customer journeys. A more mature discipline of analytical techniques for customer journey mapping is the field of process mining and discovery (van der Aalst, 2016). The idea here is to start from an event log of activities as depicted in Table 1. The event log depicts a unique customer identifier along with the various activity names and timestamps. Process discovery techniques such as HeuristicsMiner (De Weerdt et al., 2012) or Fodina (vanden Broucke et al., 2017) can then be used to discover the underlying process model or customer journey.

Various challenges arise when doing customer journey analytics. First of all, it is important that all events are properly tracked across the various customer touchpoints using a solution such as JavaScript plug-ins. It is recommended to capture all events at the lowest level of granularity. During the analysis, the complexity can be reduced by using appropriate aggregation operations as decided in collaboration with the business user. Furthermore, a customer should be uniquely identifiable across all touchpoints so that the corresponding information can be correctly matched.

Many firms will start doing customer journey analysis by focusing on one channel only, such as the web. Using data collected through JavaScript page tagging, weblogs, and cookies, firms can do clickstream analysis and understand how customers navigate through their website. More specifically, they can find out where customers come from, what search engines and search terms they used (if any), what their first page was (also called landing page), what the other pages in their session or visit were, and where they dropped out and at what time. As another example, consider an online sales process which may consist of the following steps: add items to basket, check out, provide personal details, provide payment details, review order, and confirm order. Customer journey analysis can be used to analyze how long customers spend in each of these stages and where and when they drop out. The analysis can be further enriched by performing segmentation and considering how different segments of customers—in terms of geographic region, referrer, or gender (if available)—may have different customer journeys.

References

De Weerdt J., M. De Backer, J. Vanthienen, B. Baesens. “A Multidimensional Quality Assessment of State-of-the-Art Process Discovery Algorithms Using Real-Life Event Logs,” Information Systems, Volume 37, Issue 7, November 2012, pp. 654–676.

Richardson, A. “Using Customer Journey Maps to Improve Customer Experience,” Harvard Business Review, Nov. 15, 2010; www.hbr.org/2010/11/using-customer-journey-maps-to.

vanden Broucke, S. K. L. M., J. De Weerdt. “Fodina: A Robust and Flexible Heuristic Process Discovery Technique”; www.processmining.be/fodina, 2017.



Newsletters

Subscribe to Big Data Quarterly E-Edition