In this issue, I continue with the value proposition and architecture concerns in the capture and integration of major customer interaction systems into the CRM-ready data warehouse with a look at Web clickstream data.

Though we no longer see the "e-" preceding every software, service and product offering as it has largely been relegated to a channel that doesn't cannibalize investment in legacy customer interaction channels, it's still a required competitive investment. For large organizations, it could easily cost millions of dollars per year to operate and maintain their e-presence. Is this money well spent? Does the layout facilitate revenue generation and does it do so as efficiently as possible? Are we collecting actionable customer information? What are our best e-campaigns? How do we personalize the user's experience?

To stand out, the e-channel must be utilized to the fullest to reach its potential and become a means of achieving customer intimacy. Nowhere in an enterprise will more data be generated than in an actively used e- environment, whether that environment be B2B or B2C. Also, nowhere is there more untapped and misused potential for achieving customer intimacy than in the analysis of e-data. The room for error is tremendous, especially if yours is one of the 95 percent of large-company environments that utilize dynamic content servers that can easily disassociate URL names from the pages they reference.

If you do accept the clickstream challenge, keep in mind that simple solutions are incapable of delivering in this environment. Reports documenting simple statistics such as page hits and views that are outside the context of the viewer cannot be correlated to company CRM goals. Deciphering the formats, capturing the volume and creating actions can be an immense challenge. Today, there are various vendor solutions in the market. There are also homegrown approaches to e-metrics. The technical architecture is the key to the ability to deliver a comprehensive e-metrics solution. Five key considerations drive the clickstream data warehouse architecture.

Scalability. Scalability in the context of e-metrics means the solution's architecture must be able to accept new hardware and software components such as disks, CPUs and memory; however, the solution must acquire these new components very judiciously and make the full use of them as they are added. "Hits per day" is a common means to determine the scalability of an e-metrics solution. A highly used e-channel can have hundreds of concurrent users, up to one billion hits per day and make use of up to 64 processors.

Manageability. Overworked information technology (IT) organizations need operability as close to plug-and-play as possible. Vendor solutions should offer built-in analytics that are actionable without copious customization and high headcount requirements.

Openness and Flexibility. The technical components of a packaged e-analytics solution should be as feature-rich as standalone components performing roles that are fit-for-purpose in the environment. Furthermore, they should be compatible with all components of an existing technical infrastructure. The closer the alignment, the less internal support of the technology will be required. Graphical, exception-based displays are required for analysis of e-data.

Performance. With massive volumes of data to sift through, parallelism is required in the solution. An architecture must also not easily get "backed up" by the constant inflow of data and must analyze the results and generate the reports and accessible environments in "e-time." Data collection should be nonintrusive and not interfere with the operational e-environment. Continual collection of new e-data must occur even in the event the solution temporarily becomes unable to transmit the data to the clickstream data mart.

Data Accuracy and Currency. Expectations for customized Web experiences abound; and, in e-time, there is not much slack granted for long business cycles. Data must flow rapidly through the architecture to the points of analysis and distribution. Accurate data makes all the difference in data analysis. Knowing how to clean e-based data is a developed rigor. The architecture must find threads through the data that are correct and associate e-data to all of its dimensional properties in order to make it suitable for trend analysis, customer analytics and marketing campaigns.

The filtering of unactionable Web-generated data and meaningless robotic data is a primary function of the e-analytics solution. A subject-oriented data model helps ensure that data has business meaning.

The clickstream challenge is due to large amounts of data and complex structures and relationships amongst data. Regardless, e-metrics are worth pursuing in order to optimize the e-channel for customer focus. In this pocket of CRM data warehousing, the buy value proposition is very strong today and vendor offerings should be considered for their ability to deliver across the key considerations while fitting into the overall CRM-ready data warehouse architecture.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access