The Office for National Statistics reported a 2% contraction in GDP in the fourth quarter of 2011. Furthermore, if we are to believe the Confederation of British Industries, the usual 2102 post Christmas slump in retail sales was deeper than usual with the sector seeing the biggest aggregate fall in sales volumes since the last recession. There does not appear to be respite in sight as family budgets are under continuing pressure with wages stagnant.

Retailers, usually the first to suffer in a downturn, have to look for new ways to invigorate their market and gain sustainable advantage over their competitors. One option is the creation and use of some of their vast amounts of customer data that are already available, and becoming easier to collate as technology provides different ways of capturing people’s behaviour.

The answer is retail analytics . Retail analytics advocates the use of data to get a better understanding of customers and, crucially, predict their behaviour, help retail companies to differentiate themselves from competitors, and fuel growth in challenging markets. The wealth of data available to retailers presents the opportunity for competitive advantage, but only for those who exploit it most effectively. There is no point generating so much data and allowing it to lie fallow.

Retail analytics has at its core the process of taking data from multiple sources of the retailer’s business and translating this intelligence into performance improvement. Historically, these data sources have included loyalty programmes, point of sales terminals, market surveys and market research. And some of the rewards have been enormous. Take Tesco for example: they used the date from their Clubcard loyalty programme to build unique and insightful customer profiles, to the point where it now has 12 million unique profiles for its 15 million customers (Competing on Analytics: The New Science of Winning).

These data sources will continue to retain their relevance as customer retention initiatives could be the difference between keeping customers happy or losing customers. The current challenging environment and increasing competition, however, mean there is need to progress into further areas and drawing on other inferences. A retailer can achieve growth by being able to predict the behaviours of their own loyal customers; even more growth is achievable by being able to predict the behaviours of the customers they do not currently have.

Breaking new ground in gathering useful information will inevitably involve internet-driven data gathering. Web browsing histories and online purchasing habits have become an invaluable source of data. The likes of Amazon use these data sources to support their recommendation features to their customers, while other retailers undertake targeted advertising using consumer purchasing patterns and information gathered on the internet.

Extracting datasets from social networking sites and fusing those with data already available will help retail analytics break new ground. Facebook and Twitter have together over 1.1 billion users (Facebook 845m, Twitter 300m), with Facebook reporting that their average user spends 6 hours 51 minutes per month on the site. This is a data minefield for the evolution of retail analytics and taking advantage of all that data to operate and manage the business better requires analytics.

Online retailers have shown dexterity in targeted promotions and awareness based on most recent purchase. For example, a customer buying a golf club is targeted with additional golf equipment, perhaps at a discount to reflect his loyalty, before exiting the check out.
It is in this area that in-store analytics may be behind the curve. A customer will rarely receive any promotions based on her shopping basket and will not be targeted before leaving the store. However, with better real-time availability of shopping information, it is possible to offer products and promotions on the spot. For example, certain product purchases are supplementary and could well be made together, though the purchaser may not realise it yet, and letting her walk out of the store may mean the business goes elsewhere.

Examples of Winning with Retail Analytics (source: Babson Executive Education- Realizing the Potential of Retail Analytics – Plenty of Food for Those with the Appetite)

Tesco has profited greatly from the introduction of its loyalty card program ClubCard. 80% of Tesco’s sales can be tracked through ClubCard providing rebates of 1% of customer purchases. In addition to the rebate, customized coupons based on shopper behaviour are provided to customers.
Online retailer has an aggressive program to identify and prevent credit card fraud, which in its first six months led to 50% reductions in fraud.

Amazon uses a scoring approach to identify the most likely fraud situations in customer purchases. Some of the circumstances conducive to fraud include purchases of easily-resold goods on the gray market (such as electronics), the use of different billing and shipping addresses, and use of the fastest shipping option. Such variables are used to identify and prioritize cases for investigation.

Waitrose has developed a new system for store-level sales and demand forecasting. It takes into account holidays (including Pancake Day), promotions, and seasonality for predicting demand and feeding replenishment processes. Forecasts are produced for each category and SKU per day per store. Benefits from the new system include more efficient inventory levels, particularly for refrigerated products; improved accuracy in replenishment of dry goods; a 40% reduction in order changes, and less time spent by managers in forecasting, freeing them up for customer interaction.

Online retailer used marketing mix models and test market advertising to determine that broadcast television advertising was not cost-effective. The ads did increase sales, but not as much as other marketing approaches, including offering free shipping. Amazon stopped TV ads altogether and dismantled its five-person TV advertising department.