Times are tough for retailers; they must work increasingly hard to win and retain customers. This week has seen Clinton Cards go into administration with like for like sales falling by 3.3% and losses of £3.7m for the six months to the end of January.
Clinton Cards falling footfall has been a significant factor in them arriving at this point. A lot of large retailers such as Amazon and Tesco have large banks of customer data that they can leverage to successfully target their customers with products, offers and services. Clinton Cards, and other predominantly ‘high street’ firms like them, are unlikely to possess the same quality or quantity of data – so how can they better understand, target and retain customers based on the data available to them?
For companies like Amazon and Tesco data comes from a variety of sources including loyalty card data, internet search history etc. This provides rich information about customer segments and what products and services they look for. Clinton Cards’ own data is limited to information from small, over the counter sales, without a loyalty card.
But data is available to Clinton Cards, and other retailers like them, that can provide useful information:
- Finance and performance data from 628 Clinton shops and 139 Birthdays stores
- Point of sales customer data e.g. transaction details, credit/debit card data
- Point of sales survey and other market research data
Forecasting Retail Sales
Clinton Cards offers a range of seasonal products which rely on accurate demand forecasts. Over-forecasts carry a cost in wasted products, and under-forecasts a cost in lost sales. This, combined with Clinton’s failure to predict the decline in footfall to UK high street card shops, has hit its profits hard.
One solution delivered by our Business Analytics team clusters retailers using their overall sales patterns. A forecasting model is built for each cluster using multiple linear regression techniques. The models use the most recent sales and promotional data to produce retailer level forecasts for future product sales, including the impact of any promotions.
This approach helped a major newspaper retailer to increase sales and reduce costs by understanding the drivers of demand including long term trends, periodicity, holidays, and the impact of promotions.
Despite having limited customer information to work from it is still possible for companies to understand what customers want and what drives profitability in their stores. Retailers can compare the relative performance of their stores to identify where to focus their efforts so they are less likely to miss out on major opportunities, e.g. personalised cards.
Stores are split into clusters in a similar way to the approach for forecasting sales. This time regression techniques are used to determine the correlation between store demographics, products, services, and performance data. So without detailed customer data it is possible to identify what goods and services the customers in each type of store cluster are looking to purchase.
This approach helped a major shoe retailer understand which products and services they should concentrate their efforts on to achieve their targeted £8m increase in profits.
- Many retailers are suffering as a result of the poor economy. Looking at leading indicators of how the economy may impact on customer behaviour helps develop strategic initiatives to retain those customers and maximise profits.
- Data relating to the different elements of a business can be obtained from multiple external sources (e.g. suppliers, market research etc) to extrapolate a credible model of the business. This is used to test the impact different products and services on sales revenues and profitability.
- Clinton Cards brand strength has weakened and their market share diminished. Although retailers increasingly rely on sources like loyalty cards, analysis of large customer surveys can still help to identify how their products were being perceived against competitors.