From the day of its launch in February 1995 the Tesco Clubcard was immediately embraced by customers attracted to the 1% discount off their shopping bills. But its long term success has not been built on discounts alone, rather on the personalisation of the shopping experience enjoyed by each customer through the insight gained from their shopping baskets.
To deliver this, dunnhumby has had to constantly develop the analytical techniques it uses, against a backdrop of rising cardholders and Tesco increasingly seeking to satisfy the individual needs and desires of its shoppers – among which are its best customers.
Bringing personalisation to millions
In the early days we were only able to analyse 1% of the Clubcard shopping baskets but as the cardholder numbers have grown to nearly 16 million, our techniques of interpreting the data have become increasingly sophisticated. When combined with developments in computing power it has been possible for us to analyse many millions of transactions each day.
Today, dunnhumby analyses a remarkable five billion pieces of information every week from Tesco shoppers. Every customer’s shopping basket is analysed by scoring each product against 50 different dimensions – these include for example, whether the product is foreign, branded, economy or family.
Segmenting products and customers
These products and insights are then fed into a clustering model, which splits customers into six segments: price-sensitive, traditional, convenience, mainstream, finer foods and kid’s choice. We then use an algorithm, called the ’Rolling Ball’, which ascertains links and common patterns between different products.
For instance, Alphabetti Spaghetti would be classified as a ’family’ product and by knowing that it often appears in a basket alongside a breakfast cereal like Coco Pops, we would then assign a stronger ’family’ rating to Coco Pops.
A benefit of such clustering is that we don’t need to individually classify each of the 65,000 different products typical found in a supermarket. Instead the algorithm focuses on several thousand ’seed’ products.
This is where it gets more complicated. Over time, billions of associations are created and the clustering gets ever deeper. The complexity further increases as new products are brought into the model because this can lead to the initial starting points for the analysis – the ’seeds’ – becoming less relevant and sometimes moving position.
Creating customer DNA
The next bit is the key. From the correlations and classifications of the many products in each customer’s shopping baskets we gain a complete view of the customer’s shopping preferences – their retail DNA. It is based almost entirely on the rich behavioural data as Tesco actually holds little personal information – which is little different from a shopper gazing into the trolley of the customer ahead of them in the queue.
As Tesco has used this insight to listen to its customers and respond to their needs – launching new ranges, changing store formats, sending customers’ relevant vouchers and offering new services like banking – shoppers have come to value their Clubcard very highly. It is viewed as far from intrusive.
So highly is it valued in fact, other retailers around the world recognised the benefits and dunnhumby today works with the likes of Kroger, Casino and Gruppo Pam, analysing the shopping behaviour of approaching half a billion households.
Continually developing analytical techniques
But just as in the past dunnhumby recognises the need to continually develop its techniques and not stand still. We reckon the knowledge stored on loyalty cards could further revolutionise the shopping experience in supermarkets through the further advance of personalisation.
Moving into the future – and taking one step on from the Tesco quarterly statement and its personalised offers – it is not too far-fetched to visualise screens on trolleys that alert customers to relevant offers and to then help them navigate around stores. The screens could also highlight if customers’ are buying certain products, say with high saturated fats or fatty acids.
Loyalty cards evolve
The loyalty card itself could also be in for big changes. They have to some extent evolved over time, with both Tesco and Kroger offering customers key fob variants of their cards, but (forgive the pun) much more is on the cards.
In markets like Japan and South Korea consumers are already paying for goods using their mobile phones and Tesco recently launched Clubcard on the iPhone. This will mean that even impulse and convenience purchases – from chewing gum, to a bottle of milk, to a newspaper – will also be tracked.
This extended personalisation of the shopping experience could result in the paying for goods using only our fingerprints. Trials of biometric payment systems based on finger authenticated technology have taken place in Japan and in the US, supermarket Piggly Wiggly has utilised this technology to launch a loyalty card based on a customer’s thumb print.
Linking technology and meaningful data
Although Asian markets are technologically advanced, they haven’t yet started to link the new technologies with customer data. The challenge in the future will be around managing the link between new devices and the collecting of meaningful data. As long as the customer’s interests always come first in their development, there is enormous potential in new technologies.
The likes of Tesco, Kroger, Casino, Gruppo Pam and The Home Depot are undoubtedly pioneers in the world of customer data – respecting their customers’ privacy while also rewarding them for their data. Our prediction is that if other companies follow their lead and are all collectively able to combine the emerging technologies with collecting customer data then this will herald a golden age in marketing.