Customer insight is a widely used term — but also a misused one. This article considers the core components, both defining and technical, for a customer insight function that will lead to genuine business performance improvements.
The real meaning of customer insight
‘Customer insight’ is one of those terms in the marketing lexicon that is particularly vulnerable to multiple interpretations or misinterpretation. Akin to ‘proposition’ (or marketing mix1) — which some marketers treat as if it meant product while others use it as an alternative word for promotion — customer insight is in danger of meaning all things to all people. I have heard different marketing leaders use customer insight as equivalent to just research, analytics, data or database marketing. Few outside those leading customer insight functions appear to recognize its multidisciplinary or hybrid nature.
Language is often shaped by environment, so we should not be surprised that sectors that have historically struggled to capture much data on their end customers, like FMCG, tend to use customer insight to mean research. Data-rich sectors like retail, banking or telecommunications are more likely to understand the term as relating to the use of data and/or predictive analytics. Sales-focused direct marketing businesses, meanwhile, tend to expect customer insight to mean what they value most — database marketing. So the very diversity of businesses needing customer insight is perhaps the reason for ambiguity in its definition.
Surprisingly for a discipline that has been practised in business for over 20 years, there is still no consistent definition. Even marketing gurus like Philip Kotler, Don Peppers and David Jobber avoid defining the term and instead reference component parts, like data, analysis and research. To clarify what I mean by customer insight, based on over 13 years of creating and leading such teams, I offer this definition:
‘A non-obvious understanding of your customers which, if acted upon, has the potential to change their behaviour for mutual benefit’.
When training customer insight analysts of all flavours, I stress four parts of that definition:
Insight is ‘non-obvious’ so it does not normally come from just one source of information — often it does not come from just analysis or research either. Rather, there is a need to converge evidence to glean insights.
True insights need to be actionable — hypotheses that stay theoretical and cannot be tested in practice are not insights.
Customer insights should be powerful enough that, when they are acted upon, they can persuade individuals to change their behaviour. Just benefiting from targeting based on analysing past behaviour and assuming people will be creatures of habit does not reveal any depth of understanding about them, let alone insight.
To be sustainable, the goal of such customer change must be for mutual benefit. As Peppers2 argues, a key law for marketing today is ‘earn and keep the trust of your customers’, which is achieved by acting in their best interests as well as generating long-term value for the organization.
There are also four critical technical areas that go to make up a holistic definition of customer insight. A brief review of each technical specialism may help clarify why their respective contributions are so vital:
Customer data management is a foundational requirement of the work of all the other teams. Following the old IT maxim of ‘garbage in, garbage out’, it should be obvious that the quality of any behavioural analysis, predictive analytics or targeting is dependent on the use of good-quality data. It is perhaps less obvious that this also impacts the work of much of today’s research. From targeted recruitment samples, especially when using internal segmentations, to data-capture for follow-up on Customer Effort or Net Promoter Scores (NPS), data quality is vital too. Given this, it is concerning that data teams are too often viewed and treated as the ‘Cinderella service3’ within customer insight functions.
The focus of recent hype around big data and predictive analytics has more often been analytics teams. Data scientist roles, although claimed to require more IT skills, overlap strongly with the mixture of SQL programming and statistical skills required in these teams. Behind the hype — and long before it — the role of these teams was essential to understanding how your customers are behaving and how they might behave in future. Demographic and behavioural profiling are still very important to increase customer understanding, complimented by segmentation where appropriate. But forecasting, identification of triggers and predictive models enable targeted actions to improve customer engagement and value share in future. I have spoken at a number of analytics conferences this year, and noticed that this skills set is in greater demand than ever. I hear many businesses now worrying about recruitment, retention or outsourcing of skilled analysts, even if they don’t yet have a clear plan for application areas.
Such recent enthusiasm for data and analytics should not cause us to forget the crucial role that intelligent consumer research has played in the maturity of marketing over the last 50 years. Although there is great power in being able to spot patterns in the behaviour of customers, or to use data to predict their response to trigger events, without an understanding of customers’ perception of their behaviour and that of the organization, there is a significant risk of misinterpretation. In other words, it is not enough to know how customers behave; one also needs to know why. At this point, you might rightly raise lessons learnt from behavioural psychology as to the irrational biases in how we all make decisions4 and thus the unreliability of ‘self-reporting’ in predicting behaviour. This is an important caveat on the design and use of research, as most researchers understand since their own discipline was birthed from psychology. However, companies can get it badly wrong if they ignore customers’ perception of their behaviour as well as how they may actually be making choices. Effective marketing communication requires both accurate targeting and a design that engages a customer accessibly and emotionally.
Although sometimes viewed as the preserve of direct marketing businesses, with their long history of enriching direct mail and catalogue companies, database marketing is also a vital skill for customer insight. Even if converged analysis and research can reveal an accurate understanding of customers and offers that would be welcomed at the right time, without a robust way of testing such a hypothesis it remains theoretical. Beyond that, it is very rare for customer understanding or even predictive models to get it right the first time. As with most change programmes, a culture of test-and-learn is needed to test actions that may work and refine them until an optimal return is generated. Database marketing brings the scientific method to bear on customer insight work. Effective control groups, feedback loops, statistical significance and measurement ensure an understanding of marketing payback, as evidenced by Shaw and Merrick5. As most businesses now operate multi-channel with through-the-line media mixes, this discipline has expanded to benefit from econometrics and other more complex measurement techniques.
One of Aristotle’s many legacies is his principle that ‘the sum is greater than the parts6’ — this is very applicable to the work of generating customer insights. Each of the technical areas I outlined above has a part to play. However, the value to be gained by orchestrating their combined use far outweighs what each part brings alone. I have seen evidence of this while developing appropriate segmentations, generating insights for proposition development and diagnosing actions to improve falling NPS.
To illustrate this, I use the analogy of an engine (see Figure 1). Here, customer data is at the centre. Ensuring data quality and a permission-based approach to data usage is key to its smooth working, as is some form of a single customer view. Building on this central cog, analytics, research and database marketing teams each need to operate effectively on this data. Each needs to be continuously improving in line with best practice for its discipline.
The real ‘magic’ comes from how all these four ‘cogs’ can operate together to drive action. Specific combinations vary according to the business challenge or application area. As a high-level principle, however, I have represented this in three stages of joined-up working between these technical disciplines to drive insights that can be acted upon.
At the top of the diagram is the stage of converging evidence. Here, behavioural analysts and researchers pool the evidence they have, usually defined by product/channel/segment or with relevance to a specific business issue. Through effective summarization and comparison, common themes emerge and can be used to generate an understanding of how and why certain behaviours are apparent. These can be shaped into hypotheses about customers, which can be used to propose how such customers would act in response to certain triggers or changes in communication/service/product.
The next arrow indicates applying such a targeted hypothesis through engaging database marketing. Variants of the proposition should form a test construct for database marketing experimentation. This is an iterative process until a response can be evidenced that indicates that a hypothesis has resulted in a significant change.
Finally, research is engaged once again to complement the analytical evidence from testing the experience for customers. So the customer insight team is informed not only as to whether a business change has driven a desired behaviour, but also how customers perceived this. That then naturally leads on to converging the analytical and research evidence from real-world testing, as in the first step. This ‘virtuous cycle’ of insight generation, testing and measurement continues, like moving engine parts. Akin to an engine, the point of this activity is to drive action by the wider organization (the rest of the vehicle), leading to improved commercial returns and better customer experiences (progress in the wider world).
Such a holistic model of customer insight has tremendous potential. In my own professional experience, I have seen it generate multi-million pound improvements in bottom line results. Perhaps this goes some way to explaining the recent phenomenon of more senior customer insight leadership roles within blue-chip companies7. To sustain such an influence on both marketing practice and boardroom decisions, however, I believe customer insight needs to mature as a profession and be recognized as such. Even the much-maligned profession of executive coaching now has professional bodies, with published codes of ethics and postgraduate-level courses. Hopefully, this will also emerge for customer insight as barriers are broken down between data, research and analytics professionals. This is a key challenge for customer insight leaders today.
Jobber, D. (2007) Principles and Practice of Marketing, 5th edn. McGraw-Hill, Maidenhead.
Peppers, D. and Rogers, M. (2008) Rules to Break and Laws to Follow, John Wiley & Sons, New Jersey.
Laughlin, P. (2014) ‘Don’t turn your data team into Cinderella’. DataIQ Magazine, Autumn, p20.
Kahneman, D. (2012) Thinking Fast & Slow, Farrar, Straus and Giroux, New York.
Shaw, R. and Merrick, D. (2005) Marketing Payback, Pearson Education, Harlow.
Aristotle (1998) Metaphysics, Book Η 1045a 8–10. Penguin, London.
Laughlin, P. (2014) ‘How can you influence the top table’, available at http://customerinsightleader.com/opinion/can-influence-top-table/, accessed 1 October 2014.
About this article
Cite this article
Laughlin, P. Holistic customer insight as an engine of growth. J Direct Data Digit Mark Pract 16, 75–79 (2014). https://doi.org/10.1057/dddmp.2014.59
- customer insight
- market research
- database marketing