Predicting customer behaviour is a science

Web-based businesses face similar customer issues as traditional businesses, including retaining profitable customers, attracting more loyal customers, running successful marketing campaigns and selling more products or services.

In addition, online companies, like their traditional counterparts, must meet revenue targets, customer loyalty goals and new sales numbers. Key to meeting these objectives is an understanding of the drivers behind them, and the ability to accurately predict customer behaviour.

The good news for Web-based organisations is the advent of SAS Web Analytics, sophisticated customer intelligence software enabling them to successfully address these issues.

SAS Web Analytics assists organisations to achieve increased levels of customer satisfaction, optimal use of resources and faster pay off by ensuring that the organisation really understands its customers.

"By using the correct metrics, and understanding exactly how they relate to business returns, companies can meet both strategic corporate and customer satisfaction goals," says Retha Keyser, Product Manager - Customer Intelligence, at SAS Institute, leaders in business intelligence.

"Too often companies measure certain factors, like customer retention, but do not understand the key drivers that increase or decrease retention. These may include a combination of various factors such as the number of visits a customer makes to the Web site within a determined time period, the number of products offered, the amount of elapsed time between visit, etc. Analytics can provide the answers to these questions."

SAS Web Analytics is also used to determine which new product offerings should be developed, what revenues they would potentially generate and which customers or prospects to target with these offers. Banks, for example, because they deal in virtual products, can combine products to suit specific customers segments with common characteristics. Web sites can be personalised to enhance customer satisfaction, keep them on-site, and direct them to the pay-off point as quickly as possible.

SAS Web Analytics provides insight into why people abandon their online experience. Is the response too slow? Can they not find what they are looking for because the search engine does not cater for the particular keywords customers use? Or the product is not listed in the category that they are expecting to find it under.

"In essence, SAS Web Analytics enables organisations to base decisions on facts, thus improving the online experience for customers," says Keyser. "This in turn leads to improved customer satisfaction, retention and profitability."

She explains that while most online business channels do some analysis of Web site usage, this tends to focus primarily on site statistics like session counts, page views and visitor numbers.

"None of these metrics can intelligently assist in achieving business targets," she says. "The real questions online businesses should examine are ones like: 'If I increase new visitors to the site by x percent, how will this affect my acquisition rate?'

Similarly, measurement of the Web data in isolation cannot deliver the fuller picture. For example, a delay in the delivery of an online purchase may have lead to a customer not re-purchasing from that company again. SAS Web Analytics allows for the integration of all diverse formats of online Web data with the rest of the organisation's data from areas such as logistics, sales, customer services, other channels, etc; bringing the data such as the delay in delivery into the analytical mix, so that the full picture is available and all the various business drivers are evaluated for their influence on a particular business situation.

Many companies rely on the IT department to deliver answers to their customers' Web interactions, but often IT does not have tools sophisticated or easy enough to go beyond the scope of the standard reports they produce. And if IT does have the tools, they are often not flexible enough and the delay in obtaining information to support a particular business decision may result in lost opportunity for business. SAS Web Analytics enable decision-makers to ask their own questions, and get immediate answers, freeing up the IT resources to focus on making additional data sources available to the users and using the analytical insight to drive enhancements to the Web site design to obtain improved business results. In addition, SAS Web Analytics also gives the IT department accurate daily forecasts of site statistics, which allows them to optimise expensive IT resources, increase the accuracy of capacity planning and maintain the service levels that are required.

SAS Web Analytics adds value to online channels by not only providing standard reporting but also descriptive statistics and predictive modelling.

At the lower end, it provides standard reports about, for example, visitor and session counts, as well as interactive and dynamic reporting enabling users to drill through information. However, this type of reporting still only represents what happened in the past, allowing business users to compare, for example, how the number of visitors to the site last Sunday evening differed from those who came online on Monday morning.

"This kind of information, however, cannot indicate what customer behaviour is likely to be like in the future," says Keyser. "What online businesses really need to know is which of the customers that will be coming online are likely to purchase a particular product and once they are accessing the website, how to progress them to the pay-off point in the most efficient manner."

Using advanced statistical techniques, organisations can profile and segment their customers, which gives them more insight. SAS Web Analytics identifies customer segments by automatically determining the key differentiators that distinguishes one group from the next. This gives business the ability to align marketing objectives with a particular customer segment.

"Advanced analytics detects causal drivers behind problems and opportunities, which can be used to improve the business," says Keyser.

It is therefore SAS's predictive modelling capabilities that give really valuable returns. Historical information is used to build predictive models that enable new customers to be scored individually: for example whether they have a high, medium or low propensity to purchase a particular product using the particular drivers for purchase as inputs into the model.

"Combining this intelligence with profiling information, predictive models are used to improve and refine marketing strategies leading to improved campaign response rates by delivering smaller, more focused campaigns. This in turn will lead to an increase in campaign returns and a reduction in the cost of marketing, by only marketing to those customers who are highly likely to take up a particular offer. Which will lead to a reduction in the customer irritation factor and an increase in customer satisfaction, as customers will no longer be bombarded with irrelevant and inappropriate offers." says Keyser.

It is also important to track the response and feedback from customers that have been made offers using predictive models, as this would allow one to not only measure the accuracy and performance of these models, but also to improve and adjust them as customer behaviour changes.

SAS Web Analysis therefore goes far beyond reporting, to answer questions from: 'Who is visiting?' and 'Who is buying products?' to 'When is a customer in danger of leaving the site, or not doing business there again?' SAS Web Analytics thus delivers answers to a broad range of questions to assist business in deciding which Web designs and campaign strategies are most effective in increasing not only the number of visits to the Web site, but also converting them to purchases and re-purchases in order to achieve their strategic corporate and customer loyalty goals.