Big data – In search of data gold

Potential for retail

Various industries such as insurance companies and banks, automobile manufacturers, energy utilities and the consumer goods industry have begun to use big data applications for their business in pilot projects and approaches. The retail sector likewise has high hopes for the targeted evaluation of data. In the area of marketing in particular, businesses were quick to recognise that digital information could be turned into added value.

The online mail-order company Amazon is a pioneer in this field. It is able to make accurate purchasing recommendations to its customers based on their searches and by intelligently combining this with other information taken from, for example, social networks. Here’s how it works: an Amazon customer searches for a book on the website. In addition to being shown the standard information about the item, they are also told by the website that their brother and their best friend enjoyed the book. This information is taken from the customer’s social network, where they have “liked” the Amazon page. In other words, Amazon uses its customers’ networks to make its services, recommendations and advertising promotions even more personalised.

The US retailer Walmart uses a similar system. The company collates and processes millions of pieces of data an hour, relating to, for example, how much somebody buys, when and how often they shop, and how much they spend per store visit. It then combines this with information about its customers taken from other sources, for example social networks such as Facebook and Twitter. Walmart is therefore able not only to remind customers of their friends’ or relatives’ birthdays but also to give them some ideas for presents.

The mail-order company Otto uses its data to improve supply planning for its entire product range. To this end, the company has implemented software which can more accurately forecast sales, right down to specific items. Up to 300 million data records are fed into the system each week. Important factors include the degree of advertising of an item, specific product characteristics and market conditions. Otto uses this information to generate a billion individual forecasts a year for the sales of items over the next few days and weeks. According to information provided by Otto, the mail-order company consequently has much less surplus stock and makes savings of double-digit millions of euros.

The drugstore chain dm is an active user of big data solutions too. It calculates its provisional daily sales figures based on analysis of its sales patterns over several years, together with other information such as weather forecasts. In this way, dm is able to compute its staff requirements per store and plan its employees’ working hours between four and eight weeks in advance. As a result, last-minute staff planning changes have become rare at dm.