Michael Whitelegge from M&S and Professor Mark Birkin from Leeds University hosted a thought-provoking lecture on the big data phenomenon at Leeds Business School this week. Contrary to many talks on the subject the pair focussed on practical examples: what organisations are doing and what real world impact has been achieved.
The event was tied to the recent announcement that Leeds University’s Consumer Data Research Centre has just been awarded £5m from the Economic and Social Research Council in conjunction with UCL. The CDRC aims to be “a national resource that will make data, routinely collected by business and local government organisations, accessible for academics in order to undertake important research in the social sciences to inform policy development, implementation and evaluation.” The government has identified big data analysis as one of “eight great technologies” in which the UK is internationally competitive.
Professor Mark Birkin introduced the topic, and outlined the CDRC’s remit. By gaining access to industry datasets the team hope to examine hypotheses and see what commercial value can be extracted in fields as diverse as transport, consumer trends, shopping habits and health.
Michael Whitelegge described M&S’ approach to big data, which he sees as just “more data”. His definition made reference to Gartner’s 5Vs:
- Volume – for example the data trail left by 21 million customer visits to M&S each week. Much of this will be unstructured data which requires new technology and skills to examine;
- Variety – a rich data palette, much of which is operationally generated, e.g. from credit card usage, RFID chips embedded in stock, even data feeds from chillers in store;
- Velocity – online retail in particular generates real-time opportunities, e.g. offers/vouchers based on historic purchasing offered at POP, in contrast to much customer insight which uses data “at rest”;
- Veracity – questioning the trustworthiness and robustness of analytical outputs: are correlations spurious, conclusions valid?;
- Value – the “3 whats” i) what is the data telling us? ii) so what does this mean? iii) what does the business do about it?
M&S’ aim is to develop a “data-preneurial culture” by combining analytical technology with cultural change: empowering employees to develop and test hypotheses using desktop tools. They are using Cloudera as their main technology provider. Whilst acknowledging that big data is still approaching the peak of inflated expectations on Gartner’s Hype Cycle, it is here to stay:
- It allows disparate data sources to be jammed together making analyses much more interesting. When EPOS data (what you bought, where) is merged with credit card data (which other retailers you shop at, how often) a clearer picture of the consumer emerges. Another example related to the multichannel shopper: a single customer view using credit card data means you can identify the customer who abandoned a basket at checkout online (due to delivery costs) and bought the product in store the next day;
- 80% of the data out there is new – “unstructured” data which would not have been captured prior to cheap data storage. Smartphone data showing how people move around store informing layout plans for example;
- It can create new commercial value. For example M&S’ general merchandise segmentation is based on the behaviour of 27 million customers. Targeted, relevant personalised offers result – for example a personalised clothing offer sent with a credit card statement resulted in a 40% uplift in sales.
Many of the questions after the lecture related to the ethics of examining our behaviour quite so closely. Analyses can result in profound insights which might be deemed intrusive. Whitelegge remarked that a key test for brands is not to cross “the creepy line.” Google could email you tomorrow lunchtime, saying “don’t go to the sandwich shop you normally visit when you’re working in Leeds, here’s an offer for a new joint on the high street”.
In summary? What’s possible may be unpalatable. Big data may well be a “new thing” but it, like other business operations, should be guided by an organisation’s values.