page 1
page 2
page 3
page 4
page 5
page 6
page 7
page 8
page 9
page 10
page 11
page 12
page 13
page 14
page 15
page 16
page 17
page 18
page 19
page 20
page 21
page 22
page 23
page 24
page 25
page 26
page 27
page 28
page 29
page 30
page 31
page 32
page 33
page 34
page 35
page 36
page 37
page 38
page 39
page 40
page 41
page 42
page 43
page 44
page 45
page 46
page 47
page 48
page 49
page 50
page 51
page 52
page 53
page 54
page 55
page 56
page 57
page 58
page 59
page 60
page 61
page 62
page 63
page 64
page 65
page 66
page 67
page 68
page 69
page 70
page 71
page 72
page 73
page 74
page 75
page 76
page 77
page 78
page 79
page 80

16 dunnhumby has been instrumental in the evolution of customer data. When we first began working with Tesco in 1994, dunnhumby was a young, entrepreneurial, database marketing specialist that had gained a reputation for the clever management of geodemographic data for clients like BMW and Mercury, the telecommunications provider. From the sample data of the Clubcard trial, which was running across just 14 stores, Tesco could see it had billions of items of transactional data, but it hadn't yet been able to extract profound insights from the information. Beginnings... dunnhumby immediately got to work analysing the Clubcard transactions, delivered to their offices on computer tapes. At this point in time, we were analysing around half a percent of Tesco's shopping baskets - around 500,000 different items. One of the most interesting insights to come from the analysis was that Tesco's most loyal shoppers, a small proportion, accounted for a massive part of its profitability; the 80- 20 principle. Before the data had been analysed, the economic significance of these loyal shoppers had not been appreciated. Today, around 50- 60% of Tesco's profit comes from 20% of its customers. It didn't take long for Tesco to act on this golden information and start engaging with these valuable customers. The Amersham store invited 300 customers to meet the store manager over a breakfast of orange juice and croissants. That morning, the car park was crammed with BMWs and Mercedes as some of Tesco's most affluent customers clamoured to form a closer relationship with their local supermarket. Another Tesco manager used the information to personally apologise to his customers while his store was being refitted. Before Clubcard had come onto the scene, Tesco would have put an ad in the local newspaper. Now, the store manager was able to write personal letters apologising for the inconvenience, giving advance notice of disturbance with some money- off coupons. On the morning of November 22nd 1994, Clive and Tesco Marketers Tim Mason and Grant Harrison presented the results of the Clubcard trials to the Tesco board meeting. Their argument was simple. If the Clubcard replicated the sales growth seen with any colour you like as long as it's any colour you like

17 the dunnhumby journey the trial, Tesco could become the UK's number one supermarket for the first time. The board listened to the results in silence. When the presentation was finished, after what seemed like an interminable pause, Sir Ian MacLaurin spoke from the chair: " What scares me about this is that you know more about my customers in three months, than I know in 30 years." Rolling Ball As the successful trial suggested, Tesco Clubcard was snapped up by shoppers on the day of its launch in February 1995. Two weeks later, 7m cards had been delivered to stores and 70,000 were being sent out daily to replenish stocks. Within days, over 70% of all Tesco sales were being recorded from Clubcard holders, 3.5m weekly transactions. The Clubcard hotline was taking 1,000- 2,000 calls a day from shoppers keen to use their cards to get 1% discount. In the early days, we were able to analyse just 1% of the Clubcard shopping baskets. As Clubcard members mushroomed, dunnhumby's techniques of interpreting the data to make sense of strategy increased in sophistication, while developments in computing enabled us to analyse millions upon millions of transactions. Today, dunnhumby analyses nearly 5bn ( 4.8bn) pieces of information a week from Tesco. Every customer's shopping basket is analysed by scoring each product against 50 different dimensions. Are these products foreign, branded, economy or family, for example? These products and insights are fed into a clustering model, which has identified six segments of people: price- sensitive, health- focused, traditional, convenience, mainstream and upmarket. We use an algorithm, called the Rolling Ball, which draws links and common patterns between different products. For example, Alphabetti Spaghetti would class as a ' family' product. By knowing that it often appears in a basket alongside a breakfast cereal like Coco Pops, dunnhumby would then assign a stronger ' family' rating to Coco Pops. This efficient clustering technique means that dunnhumby doesn't need to individually classify the 65,000 different products a typical supermarket carries. Instead the algorithm focuses on