From Missile Tracking to Customer Targeting
The term “customer targeting” refers to the customisation of a firm’s marketing mix to increase the appeal of the firm’s product. With the advancement in information technology, data collection and analysis have become easier.
by Chung Tuck Siong
The term "customer targeting" refers to the customisation of a firm's marketing mix to increase the appeal of the firm's product. With the advancement in information technology, data collection and analysis have become easier and this allows firms to work towards the targeting of customers on a one-to-one basis. The idea is to work towards what economists call "first-degree differentiation". Ideally, a firm can reap the most profit by selling a different product to each and every customer and charging the maximum price that they are willing to pay for the customised product. Customising products using psychographic and demographic profiles are two of these initiatives.
However, there are still challenges that firms need to overcome in their customisation efforts, namely - consumers' preferences are not always static. Context, we now know, is a major variable that affects consumers' product choice decisions. Take music for example, the type of music that an individual chooses to listen to at any point in time depends on many factors including their current location, as well as the mood and company they are in. Companies designing customised targeting campaigns therefore need to understand the relevance of context in order to create successful campaigns.
To help tackle the problem of context-based personalisation (customisation at a one-to-one level) we turn to an unlikely source for inspiration – missile tracking. Specifically we have exploited the statistical technique called "particle filtering" which has been used to help track missiles. There are many applications for this technique and most of them have to do with estimating the current spatial/geographical position of a moving target.
In our case, think of the moving target as the changing contexts in which a customer finds himself. By having a technique that automatically detects the specific customer context, marketers are better able to provide context-specific and personalised marketing mixes. For example, think of the provision of the appropriate type of music when the customer is in the privacy of her own home after a long, difficult day. Obviously some music forms are more suitable than others in providing comfort and relieving stress. This might also be a good time to promote psychographically and demographically personalised holiday packages to help lift the customer's mood. Therefore, the automatic detection of context and the context-specific personalisation of the marketing mix stand to make any product offering much more relevant to customers.
We implemented a version of the particle filtering technique in our field study. The study was carried out using undergraduates as participants each of whom were issued with a personal digital assistant (PDA). The PDAs were equipped with the personalisation software we developed which allows the personalisation of music clips presented to the participants. The participants were able to carry these PDAs around and to listen to the music in places and times of their own choosing.
At the start of the study, the participants listened to a random and diverse mix of music clips with different musical attributes. This helped us to estimate the preference range for each participant's taste in music. In regression terminology, what we did is akin to obtaining different sets of coefficients for the different musical attributes for each participant. A coefficient here is an indicator of how much importance a participant places on a particular music attribute. We can think of these different sets of coefficients to represent the preferences for participants when they are in different contexts.
Armed with these coefficients, our particle filtering technique helped zoom in to the more preferred set of songs when participants were in different contexts. After observing the listening behavior of a participant, the particle filtering portion of our software helped to adapt the personalised song to suit the existing detected music preferences. In addition, the technique allows for changes in music tastes and the seeking of variety in music clips. All that it requires is some recalibration of the sets of coefficients at suitable intervals. We demonstrate that such an approach outperforms alternative personalisation methods without the adaptation.
There are many additional factors which we are now in the process of testing in order to fully optimise the breadth and flexibility of the tool. However, our current study does herald the exciting possibility of a real time, context-based product personalisation approach that captures the micro-moments that customers are in. Sometimes solutions can be found in the most surprising places!
About the author
Chung Tuck Siong is the Assistant Research & Development Director of ACI, and Assistant Professor in Nanyang Business School's Division of Marketing and International Business. His research expertise spans the areas of services and digital marketing, and he has published, among other works, a well-cited article titled "Marketing Models of Service and Relationships" in Marketing Science, a top-tier marketing journal. He has taught Services Marketing for a few years and is now teaching Market Relationships at the Nanyang Business School. He is also the chair of the Marketing and International Business division's PhD committee.
