Analysis: CRM analytics - crunch those customers

Once you have data on customers, how do you make it sweat? Anthony Plewes looks at an aspect of customer relationship management that may finally be finding its feet.

Not long ago any discussion of data mining and CRM analytics would have made the average business executive’s eyes glaze - data mining algorithms and OLAP seemed to belong firmly in the realm of the statistician.

But no longer. CRM analytics is big business. Gartner Group has identified customer service analytics as the fastest growing customer service application through 2005. And the tools for carrying out analysis have been taken out of the lab and given directly to line-of-business managers. No longer the preserve of the specialist, CRM analytics is "modelling for the masses", quips Dr Judy Bayer, NCR Teradata EMEA director for advanced analytics.

Data mining and analytics allow companies to find patterns in large volumes of data and it is particularly useful for industries that collect large volumes of user data such as those in telecoms, financial services and retail. This data can be used to derive information, which in turn gives the company knowledge and insight into customers’ behaviour.

Analytic capability is appearing throughout the CRM value chain. In Hewson Group’s 2003 market analysis of CRM, market leader Siebel has grown its analytics division from almost nothing two years ago to over $200m in 2003. Siebel entered this space through its acquisition of nQuire.

“Analytics is proving to be very popular,” confirms Neil Morgan, European marketing director for Siebel. “It went straight to number two in our software chart and is the most popular new product.”

Many key CRM and call centre vendors are increasing their analytical capability, with products coming to market with embedded analytics and data mining.

“CRM vendors should be interested in analytics because it gives them more control in the project,” says Nick Hewson, managing director of CRM consultancy, Hewson Group. “It is a powerful position for them to be in.”

CRM analytics are not just available from CRM vendors. Other vendors include CRM analytic specialists such as Ephiphany and Quadstone; ERP vendors including SAP, Peoplesoft and Oracle; and data warehousing and analysis companies such as NCR Teradata and SPSS.

Some of the smaller vendors are likely to be acquisition targets for larger CRM vendors. “We expected there to be more consolidation in the market last year,” says Hewson. “There are some capable tools out there.”

CRM analytics covers a wide range of activity. Its primary function is to allow companies to perform segmentation on their customers.

“You can use analytics to understand and predict customer value,” explains John Radcliffe, research director at Gartner Group. “This forms the basis of customer segmentation and allows companies to allocate their resources according to customer value.”

Most of this analysis is based on historical data, such as cost of customer acquisition compared to customer value. “Doing this alone is a challenge enough for some organisations,” says Radcliffe. “Many companies can’t even measure customer profitability, let alone use data mining to predict customer behaviour.”

Of course companies can only do customer analytics if they know who their customers are. This is especially a problem if they sell through retail outlets for example. Supermarkets solved this problem using loyalty cards to help them identify customer buying patterns. Without this they are only able to carry out analytics at a market or basket level.

And because analytics gives insight into an organisation’s data, analysis will be of little use if the source data is poor. The age-old maxim of 'garbage in, garbage out' applies irrespective of any clever algorithms. A PricewaterhouseCoopers survey carried out a couple of years ago found that 75 per cent of respondents across the US, UK and Australia reported significant problems and losses as a result of defective data. Only one in three respondents were very confident in the quality of their own data. Gartner’s Radcliffe says that data quality continues to remain for many companies.

If implemented properly on clean data, the benefits of CRM analytics can be significant. Over the past few years return on investment (ROI) in CRM has received a battering, largely because companies at the height of the CRM craze rolled out software without measuring business benefits. Analytics can help companies get ROI on their CRM investment. “It is difficult to drive proper operational processes without analytics,” says Nick Hewson.

“While there is a lot of cost involved in making sure the infrastructure is up to scratch, by its very nature you are counting and measuring so you will be able to measure tangible results,” says Gartner’s Radcliffe. And although analytical CRM is harder to implement and requires a longer period of time to realise returns than operational CRM, the potential ROI of analytics will continues to grow, because organisations can continually optimise relationships.

The key trend in CRM analytics is that the tools are increasingly being designed with the business user in mind. The analysis tools have complex data mining functionality in the background but the user does not need to understand the specifics. Line-of-business managers are now able to have direct control of the analysis, rather than having to send batch data to a specialist.

“When planning a campaign, the manager needs to get the results from analysis quickly,” says Radcliffe. “However, batch processing does not allow this.” Now that the business managers are in control, marketers are able to run multiple micro campaigns to hone their strategy.

“It is possible to do clever things to data without having to extract it,” says Teradata’s Bayer. “This allows you to be more creative and do analysis faster. You can think about hypotheses and experiment with scenarios. This reflects a trend: people are beginning to focus on data rather than algorithms.”

This trend is even creating a slight reluctance to use the word ‘data mining’ as it indicates a focus on technology and algorithms, rather than business.

As analytics moves away from a batch-processing environment, more organisations are able to make use of analytical data on the fly. For example, a customer’s score can be instantly updated after an interaction. “We see the use of analytics shifting from batch to real-time,” says Michael Trigg, VP corporate marketing at Epiphany. “This allows companies to use analytics to guide their interaction with customers.”

“The business intelligence market is mature and commoditised,” says Trigg. “Most vendors offer that as part of their toolset. We find that organisations have a need for analytics at the customer interaction point.”

Epiphany’s software takes a real-time profile of a customer and decides what treatment to extend. The profile can include information such as gender, postcode and churn scores. This information is then used to calculate cross-selling and up-selling opportunities, for example.

This functionality is not just for call centres. NCR Teradata is working on a project in Europe that aims to bring real-time analysis to the screen of automated cash tellers (ATM). “When you withdraw money at the ATM there is a personal message for you, such as a marketing offer for a loan,” she explains. It has developed similar functionality that is already in production at a supermarket chain in France, where a customer is offered vouchers after shopping, based on a real-time analysis of their purchase.

Real-time functionality and analytic tools for business users are two key reasons why analytics will increasingly become a key part of all CRM implementations.