Reference no: EM13946348

Case study

Boots mine diamonds in their customer data

The high-street retailer Boots launched its Advantage loyalty card in 1997. Today, there are over 15 million card holders of whom 10 million are active. Boots describes the benefits for its card holders as follows:

The scheme offers the most generous base reward rate of all UK retailers of 4 points per £1 spent on products, with average card holders receiving 6.5 points per £1 when taking into account all other tactical points offers.

There are 23 analysts in the Customer Insight team run by Helen James who mine the data available about card users and their transactional behaviour. They use tools including MicroStrategy's DSS Agent and Andyne's GQL which are used for the majority of queries. IBM's Intelligent Miner for Data is used for more advanced data mining such as segmentation and predictive modelling. Helen James describes the benefits of data mining as follows:

From our traditional Electronic Point-of-sale data we knew what was being sold, but now [through data mining] we can determine what different groups of customers are buying and monitor their behaviour over time.

The IBM case study gives these examples of the applications of data mining:

What interests the analysts most is the behaviour of groups of customers . They are interested, for example, in the effect of Boots' marketing activity on customers - such as the impact of promotional offers on buying patterns over time. They can make a valuable input to decisions about layout, ranging and promotions by using market basket analysis to provide insight into the product purchasing reper- toires of different groups of customers.

Like others, Boots has made a feature of multi- buy promotional schemes in recent years with numerous ‘three for the price of two' and even ‘two for the price of one' offers. Using the card data the Insight team has now been able to identify four groups of promotion buyers:

¦ the deal seekers who only ever buy promotional lines;
¦ the stockpilers who buy in bulk when goods are on offer and then don't visit the store for weeks;
¦ the loyalists - existing buyers who will buy a little more of a line when it is on offer but soon revert to their usual buying patterns; and
¦ the new market - customers who start buying items when on promotion and then continue to purchase the same product once it reverts to normal price.

‘This sort of analysis helps marketeers to under- stand what they are achieving via their promotions, rather than just identifying the uplift. They can see whether they are attracting new long term business or just generating short term uplift and also the extent to which they are cannibalising existing lines', says Helen. Analysing market basket trends by shop- per over time is also providing Boots with a new view of its traditional product categories. Customers buying skin-care products, for example, often buy hair-care products as well so this is a good link to use in promotions, direct mail and in-store activity.

Other linkings which emerge from the data as Helen says, quite obvious when one thinks about them - include films and suntan lotion; sensitive skin products - be they washing up gloves, cosmetics or skin-care; and films and photograph frames with new baby products. ‘Like many large retailers we are still organised along product category lines', she says, ‘so it would never really occur to the baby products buyers to create a special offer linked to picture frames - yet these are the very thing which new parents are likely to want.

‘We're also able to see how much shoppers participate in a particular range', says Helen. ‘They may buy tooth brushes, but do they also buy toothpaste and dental floss?' It may well be more profitable to encourage existing customers to buy deeper in the range than to attract new ones.

Monitoring purchases over time is also helping to identify buying patterns which fuel further marketing effort. Disposable nappy purchases, for example, are generally limited by the number of packs a customer can carry. A shopper visiting Boots once a fortnight and buying nappies is probably buying from a number of supply sources whereas one calling at the store twice a week probably gets most of her baby's nappy needs from Boots. Encouraging the first shopper to visit more would probably also increase nappy sales. Boots combines its basic customer demographic data (data such as age, gender, number of children and postcode) with externally available data. However, according to Helen ‘the real power comes from being able to combine this with detailed purchase behaviour data - and this is now being used to fuel business
decisions outside of the marketing arena'.

An analysis of how Boots customers shop a group of stores in a particular geographical area has led to a greatly improved understanding of the role different stores play within that area and the repertoire of goods that should be offered across the stores. For example, Boots stores have typically been grouped and merchandised according to their physical size. This leads to large stores competing with smaller stores for trade in the same area. ‘We quickly learned that our most valuable customers shop across many stores in their area', says Helen, ‘and that there is a lot to be gained by managing stores as local areas and focusing on getting the overall customer offer right.'

Gaining a greater understanding of how cus- tomers shop product areas and stores offers really valuable insights.

However, as Helen says, ‘the real prize is in gain- ing a really good overall understanding of your customers'. For a retailer with a very broad customer base it is too simple just to focus all efforts on the most valuable customers. Boots build up their understanding by combining data from a number of customer dimensions: RFM (Recency, Frequency and Monetary value) analysis enhanced with profitability. This helps Boots to understand the main drivers of customer value and identify which cus- tomers they should value and retain and which could be more valuable if they focussed on them more.

Lifestage analysis - This provides insight into how a customer's value changes over their lifetime. Using it, Boots can identify which are the potentially valuable customers of the future. They can also see the point at which a customer might become less valuable and try to prevent this. It is also clear that some messages become very important at certain times (for example, vitamins to people over 35 who have realised they may not be immortal) and irrele- vant at others (what mother is concerned about cosmetics within a couple of weeks of the birth of her child?). This informs the mix of messages the customer receives, for instance via direct mail.

Attitudinal insight from market research surveys and questionnaires gives Boots an understanding of the attitudes driving the behaviour they see on their database. It is pointless directing a lot of marketing effort at people whose attitudes mean that they are unlikely to become more valuable to Boots.

This diversity of data is being used to build up a multi-dimensional picture of customers that gets to the heart of what drives customer value both today and into the future. Analysis of attitudes and customer repertoires offers Boots pointers to influencing customer value in a positive way. This understanding of customers has many applications within Boots from the way the Boots brand is communicated to specific cross-selling activities for store staff. One of the first applications of this segmentation was as a driver of the Boots relationship marketing programme enabled by the Advantage Card.

The segmentation provides a framework for relationship marketing. Specific campaigns help Boots to deliver that framework. These could encourage customers to shop along different themes - Summer holidays, Christmas shopping - and incentivise them to make a visit. They may simply raise awareness of a particular new product or service - Boots Health & Travel Cover launched in April is a good example of this. They could be an invitation to an exclusive shopping event where the customer can shop in peace and perhaps earn extra points as well. To make all this happen Boots needed a campaign management system that could involve customers in the relationship marketing programme most relevant to them. The ‘campaign management' com- ponent has been fully integrated within CDAS [Customer Data Analysis System] through a bespoke development by IBM. This means that direct market- ing analysts are able to develop their target customer profiles without having to first create a separate extract of the data and are also able to base these profiles on the full richness of informa- tion held within the database. Having defined these criteria, the system will automatically come up with a mailing list of matching card holders with no further intervention. The system not only automates the measurement of basic campaign response analysis, but also makes the list of customers actually mailed available within the analysis environment so that more sophisticated response analysis can be performed. ‘The close integration of the Campaign Management System within the analytic environment of CDAS is one of its main strengths', says Ian, ‘not only are we able to drive high response rates by tightly targeting relevant customer groups, but we are able to close the loop from initial customer analysis, through customer selection and campaign execution back to campaign response measurement and further campaign analysis.'

"When we announced the Advantage loyalty scheme we knew that the incremental sales gener- ated by it would pay for the initial investment, but that the long term value would come from the application of customer insights across the business', says Helen - ‘we are already proving that we can add significant value from doing this. But you do not obtain these benefits unless you get the base of detailed information right - and couple this with an ability to thoroughly exploit it.'

Source: Computer Weekly (2001b)