The 4 Requirements for Customer Insight

Different businesses continue to use the term “customer insight” (CI) to mean different things. Even in a recent poll of more than 100 customer insight leaders, only half considered data management or database marketing to be part of customer insight. The majority also had only research reporting into them, not analytics. Does that ring true with your role? In June, I shared a definition of customer insight that I find useful: “A non-obvious understanding about your customers, which, if acted upon, has the potential to change their behavior for mutual benefit.” I would stress four parts of that definition: First, that insights are non-obvious. They normally require the convergence of evidence from multiple sources to help spot themes, so you can then dig deeper for motivations. Second, that true insights need to be actionable. There is no point learning something unless you can change commercial results or customer experience as a result. Third, a good test of an “insight” is whether acting on it is powerful enough to change your customers’ behavior (not just data to target those you believe will act as they have in the past). Fourth, in this “Age of the Customer,” the importance of trust should mean any insight has the goal of mutual benefit for the organization and the customer. Anything else is short-term success for long-term value erosion. You may need to wear many “hats” to achieve that kind of insight production from your team. As an earlier post listed, you may need to be a data scientist, psychologist, artist, storyteller, sales coach, economist and people leader. Quite a challenge, and perhaps one reason why leaders like that are hard to find. Some people suggest that customer insight is making use of your data, with the current buzzword being “big data." But it is possible to be drowning in data and still none the wiser about your customers. Perhaps as a result, some leaders appear to equate CI with behavioral analysis and statistical modeling (with the buzzword here being “predictive analytics”). Such analytics can be very powerful, but, without an understanding of why customers are behaving as they are, it won’t pass the test of our insight definition. You may assume that CI as a term applies to research, qualitative and quantitative activities focused on that “why” question. But with the unreliability of self-reporting and the need to behaviorally test what customers actually do, to identify any behavioral economic biases at play, can this really be relied upon by itself? So, you might conclude that the only way to know the reality is to test and learn, using targeted communications and measurement from database marketing. That is also very useful, but what understanding helps create what should be tested? As is probably obvious from my definition, I would recommend that all four of those technical disciplines are needed, to create true customer insights. However, it is not just these separate parts operating effectively in isolation, but the synergies and insights that can be realized by the working together in collaboration. As I’ve heard different experts speak on this topic over the years and seen the progress of customer insight leaders in the field, it seems to me that we are talking about an ecosystem. So, the challenge for CI leaders becomes how to nurture this ecosystem, ensuring each part fulfills its potential and acts symbiotically with others to produce the healthy fruit of actionable customer insights (in a way that feels more organic than mechanistically following a set process). Ensuring a consistent source of quality data for all the technical teams is at the heart of this ecosystem. Then using that data will need to be skilled research, analytics and database marketing teams (brought together in one CI function). Real growth, however, it appears, happens when you use these parts together. For example, converging the evidence from analysis and research to produce a more robust picture of how customers are feeling and acting and why. That should enable hypotheses to be generated as to how customers would feel and act if you did something different. Offering something different (communication/experience/product) can then be tested with experimental design using database marketing skills. Once you can see any changes in customer behavior as a result, also check out the feelings of customers and observe the touch points to get a feel for their new experience. Such research and analysis output then brings us back to the stage of converging evidence and looking for themes (a virtuous circle of continuous improvement). There is more to customer insight generation than that. Perhaps I'll post another time about applications like generating insights for proposition design. But, for now, I wanted to share what is becoming a standard model for me in helping clients. Do those themes resonate with you? Any other tips you can share?