Mining gold in behaviour patterns
If you’re in the business of selling, then understanding your customers’ needs and expectations is the key to increased sales and profits.
The challenge is how to gather and use information from different sources.
Ideally, marketers want to achieve long-term sustainable relationships with customers rather than simply drive short-term acquisition initiatives. This requires a deeper understanding of both customers’ behaviour patterns and their attitudes.
Market research and data mining are essential to achieve this. Market research will allow you to understand more about your customers’ motivations, feelings and attitudes. Data mining will reveal customers’ behaviour patterns and characteristics.
The challenge then is how to integrate the two sets of information in a way that offers the most value.
Market research usually involves sampling a small percentage of customers in order to understand their psychological states. For example, Bank XYZ may survey 1000 people from its client base to learn if a proposed branch colour scheme is conducive to increased patronage. From this data, researchers make deductions about all of the banks customers. If 60% of the customers surveyed liked the new colours, that finding will be used to guide business initiatives.
An alternative market research approach is qualitative analysis. Rather than give people surveys to complete, marketers form focus groups of predetermined customer types, such as 18-25-year-old males or 65+ females. This type of research is resource-intensive and requires careful analysis to be of value. The data gathered can be used either to formulate surveys or stand on its own in helping communications specialists carry out their jobs or select issues best dealt with by data mining.
Why don’t marketers survey everybody or involve them in focus groups, removing all doubt and uncovering the opinions of all? Cost. Data mining reveals patterns in the data, such as customers’ transactional, behavioural and profile characteristics. These patterns are useful and powerful, because they serve as the basis of customer categorisation – knowing which customers have similar characteristics. This enables the creation of predictive models – predicting who is most likely to behave in a particular manner or buy a particular product.
This information is available for all customers and can be augmented with data from other external sources such as census data from Statistics New Zealand.
In contrast, data mining is cost effective, with the cost of data collection minimal. However, data mining has its hurdles, in particular the fact that customer information is usually spread throughout disparate data systems. The challenge is to link all data regarding a customer into a single customer view. A limitation less easily overcome is the lack of information pertaining to customer’ attitudes and perceptions.
The best results are achieved when market research and data mining are integrated.
The ideal scenario is to match individual customer responses from market research to their actual behaviour gained from data mining. This can provide insights into those customer’s behaviour and into that of other customers sharing similar characteristics.
Integration can help overcome the weaknesses of each method. For example, market research combined with data mining could identify patterns that enable a company to spot likely prospects or predict customer defection. Importantly, integration can be used within the confines of the market research code of practice and the Privacy Act 1993.