Listening to Customers by Triangulating Their Behaviour
Behavioural-data customer-journey methodology shipped; scalable across markets, foundation for continuous product improvement.
In brief
Situation. Customer journey mapping in most companies is done in workshops, on whiteboards, with sticky notes, by people who are not the customer. The result is a beautiful diagram of the journey the company thinks the customer takes.
Complication. Real customers do not take that journey. They wander. They abandon. They come back. They use the product in ways nobody imagined. The fixed-path journey map becomes a piece of office decor while the actual customers continue to misbehave in revealing ways.
Resolution. I co-led, with the Commercial Director of Customer Experience, a project to capture the actual journey from data, what we called the Timeline of Facts. Analytical pipelines ingested behavioural data across thousands of customer journeys in multiple markets, with algorithms determining the primary, secondary, and tertiary root causes of dissatisfaction. The two technical heads reporting to me, Big Data Infrastructure, and BI and Analytics, ran the execution.
Impact. Dramatically improved understanding of where products were failing customers. The Customer Experience team’s analytical work was streamlined and response times accelerated. The methodology is now scalable across markets and products, providing a foundation for continuous product improvement.
The longer story
The single most important question in customer experience is one most companies are afraid to ask: “why did our customer give up?” We are afraid of the answer because the answer is often “because of something we did.”
The advantage of pulling the journey from actual behavioural data, rather than from a workshop, is that the data does not lie politely. It says “on average, your customers abandon at step seven, and the reason is that step seven is annoying.”
Once we built the timeline-of-facts pipeline, every department that touched the customer had to look at the same data. The marketing team could no longer claim their campaign worked while the product team claimed the product was fine. Both teams shared a customer who was, at this exact step, deciding to give up.
Conversations changed. Defensiveness dropped. You cannot triangulate excuses; the data has only one version of events.