The next frontier in business opportunity and challenge is e-business, and hundreds of thousands of organisations around the world are rushing to embrace and maximise the opportunity.
To ensure they can address the opportunity, organisations are building powerful, foolproof, highly available e-commerce engines and gateways to financial institutions. But how much thought are they applying to the analysis of the activities of customers they`re attracting with their e-business?
This is the question posed by Simon Jeggo, GM of Ardent Software Africa. Unless they are factoring in the analysis component as a core component of their e-business initiative, they will simply be haemorrhaging money.
"Businesses must take advantage of data from their Web-based activities immediately if they are to maximise their e-commerce profits," says Jeggo. "They need to be able to respond quickly to changes in customer buying behaviour, and respond to website visitors almost immediately in order to turn them into customers."
In the past - and not that long ago - companies had the luxury of time in which to implement their marketing campaigns. Where they had days or weeks in which to measure the effectiveness of campaigns, today Web marketers need to know in hours if a campaign is effective.
"Decision support systems need to assimilate huge amounts of new data every day and they need to provide analysis in realtime, or close to realtime," says Jeggo. "This poses great technology challenges, which are solved with the technology known as clickstream analysis, or, as it`s being termed, clickstream webhousing."
For e-marketers, clickstream webhousing is an attractive concept. There is much to be gleaned from tracking how consumers make their way through a website en masse.
"The clickstream is all about working through Gigabytes or even Terabytes of customer data, as reflected by their mouse clicks, and identifying useful patterns," says Jeggo. "This is a variant of data mining."
A basic clickstream webhouse could show, for instance, that 20% of traffic leaves a website on a certain page; or that 20% of cyber visitors abandon their loaded shopping carts in a cyber-aisle, while they were busy shopping. A clickstream webhouse would show how much money was lost in these abandoned shopping carts.
"Without a clickstream webhouse, the problem could persist indefinitely," Jeggo advises.
A clickstream webhouse poses other challenges: the sheer volume of data generated in cyberspace beggars current technology`s attempts to analyse it. Leading companies such as Microsoft can generate more than a billion clicks a day on their aggregated websites around the world.
"With the best will in the world, there is no way such a volume of data can be analysed each day. Yet, if it isn`t, we`re losing an opportunity to get closer to our Web-based customers. So current data mining tools need to incorporate ultra-sophisticated sampling algorithms," says Jeggo. "The ability to sample in a manner which is accepted as statistically representative will determine the success of clickstream webhousing."
Equally important, Jeggo advises, will be the ability to incorporate Web-based data directly into existing business intelligence initiatives. "As with all new technology waves, it is imperative that the new data stream is automatically and seamlessly absorbed into the organisation`s current business intelligence platform. This allows e-commerce to feed business intelligence and vice versa, for both to drive customer relationship programmes, and all to create an iterative loop."
These requirements are driving the need for integrated data extraction, transformation and loading (ETL) tools, whose sales are reaching record levels. Ardent`s DataStage Suite, which fulfils these criteria, was bought by 451 customers globally in 1999.
"The iterative and mutually dependent nature of today`s business/technology waves, with data the common denominator, means data quality and consistency must be subsumed," concludes Jeggo.
Share
Editorial contacts