Leading retailers are utilising cutting-edge technology to get individual customer views in an instant. Retail Week finds out more.

Leading retailers are utilising cutting-edge technology to get individual customer views in an instant

Customer insight was once a behemoth. It involved tonnes of data, millions of customers and very big computers. Once the three were combined, it was slow moving. Retailers could wait weeks for any useful insight. But this approach is beginning to look outdated and cumbersome as retailers including Burberry, Debenhams and Otto make use of the latest analytics tools to make decisions on live customer data, often learning as they go.

Earlier efforts at customer insight, such as the Tesco Clubcard, relied on gathering data from customer sales and storing it in a data warehouse for analysis. Data scientists could then use analytics software to create rules predicting behaviour. They could forecast the likely response of groups of customers to product offerings, formulated in algorithms, which could then be applied to future discounts and coupons. Although effective, it required a reasonable turnaround time.

The advent of ecommerce, along with advances in computing, makes new approaches possible and increasingly necessary, says Chris Allan, managing director of products analytics, Accenture. “A lot of the old-style analytics tended to involve quite a lot of heavy lifting,” he says.

“As computing power and analytics capability has accelerated dramatically over the past few years, problems which would take weeks to run, and then be too late to be relevant to your earlier point, are now becoming very much solvable problems. People are starting to look at things about which they had once said: ‘we can’t do that’.”

This means retailers can start to get insight into individual customer behaviour, rather than in groups, and tailor marketing and special offers while the customer is present, rather than using the insight from some earlier interaction, Allan claims.

Accessing memory

One technical factor enabling this transformation is the move to performing analytics ‘in-memory’. This means analysing data in the computer’s working memory, which is much faster than recalling it from disk. Until recently this technique was so expensive it was seldom used outside scientific and financial computing.

Burberry is a chief proponent of in-memory analytics, launching a program called Customer 360, which invites customers to digitally share their buying history, shopping preferences and fashion phobias. The high fashion retailer has worked with software firm SAP to use its HANA in-memory analytics system to analyse huge amounts of data, from shoppers’ social media profiles, purchase history and other information, in real time to assess customer likes and tastes. This is then presented to sales floor staff on a tablet device to help serve the customer.

Gartner research director for information management, Roxane Edjlali, says potentially all retailers will be affected by the move to in-memory computing, often applied to real-time analytics in customer insight. This might mean sending coupons to mobile devices based on a customer’s location, preferences or time of the day.    

“There are use cases where latency is really important. If you want to have the interaction with the data that requires information to be available immediately for driving pricing or scoring algorithms, and where speed changes the way that the business process or business model works, those require hybrid in-memory transactional and analytical capabilities,” she says.

While in-memory analytics can be expensive, the advent of cloud computing means retail businesses do not have to pay upfront for cost of hardware, software licences and putting it all together. This pay-as-you go model, which hosts the system on remote servers, offers smaller businesses, with lower IT and marketing budgets, access to customer insight in real time, allowing them to adjust their offer to customers as they interact with them online.

Aberdeen-based online butcher and prepared food retailer Donald Russell once only served the top restaurants, but now it also generates £25 million revenue from consumers. It is able to design offers as they journey through its website using software developed and hosted by SmartFocus (see box).

Live and learn

These examples rely on a historic picture of the customer to create algorithms that help businesses decide how best to serve the customer based on live data. But some companies are going a step further and applying ‘self-learning’ software to live data, allowing offers to be adjusted should any new patterns in behaviour emerge.

For example, software firm Featurespace promises to help retailers spot growing trends and recognise appropriate cross-sell and up-sell opportunities based on an evolving understanding of customer intentions and behaviours.

Debenhams is applying this technology to improve customer retention by trying to find the point at which customers churn and creating engaging campaigns to keep its valued customers. It is at the early stages in deploying the technology and has declined to talk about it publicly.

German software firm Blue Yonder has also developed self-learning technology, which is being used by retailers including European online fashion giant Otto. Head of consulting at Blue Yonder, Daniel Grüssing, says the company is studying the application of live, self-learning analytics to customers as they navigate a physical store. Identified as they log in to the retailer’s wi-fi, or through their mobile phone

ID, customers receive coupons on their smartphones based on past purchasing history, time of day, store location, weather and a range of other factors.

Given the ubiquity of smartphones and the growing power of analytics software, these systems are becoming increasingly feasible. They will add to the growing number of techniques for exploiting customer insight in an instant. The question remains as to which retailers are ready to take advantage of them.

Case Study: Butcher beefs up customer insight

Donald Russell screenshot

The problem: As a butcher to the Royal household and supplier to top restaurants across Europe, Donald Russell was forced to re-evaluate its business when the BSE outbreak saw UK beef exports banned during the 1990s. A shift to supplying consumers direct via the internet followed, as did an expansion in its product range.

The solution: In such a transition, understanding consumers is paramount, says head of marketing Gary McDonald. “Being able to know what the customer is looking at, what they are buying, what they think, conversion rates and customer feedback all give a great indication of demand,” he says.

To put these customer insights at the heart of the business, two years ago Donald Russell moved to a cloud-based customer insight system from SmartFocus, which provides the ability to tailor product and promotions offered to customers in real time as they surf the food retailer’s website.

Factors influencing offers and promotions include customers’ preferences of a particular meat, whether they shop for special occasions or the weekly family groceries and their location. In all, customers are divided into 50 segments, McDonald says.

The systems can predict which online content will likely lead to higher sales and conversions rates, based on historic data of what customers have looked at and bought. Their website movements are also tracked using a cookie after customers have registered with the firm.

The results: Such insights are used to determine website content, and personalise up-sell offers and product recommendations to the segments of its best customers. Since the introduction of the technology, Donald Russell has seen increased conversion rates, a £2 increase in the average online order and a greater frequency of orders, according to McDonald.