Using Analytics to Predict Customer Demand

There's much gnashing of teeth and election-year jockeying over the decline of manufacturing or services-jobs outsourcing, but most of corporate America has already moved on to the next imperative: perfecting the demand chain--activities and processes that create, nurture, and capitalize on demand for products and services. This is the 21st-century moneymaker, and many companies are attempting to achieve perfection using CRM solutions that collect a range of customer behavior and preference data. However, to truly achieve perfection these organizations have to find a nimble way to gain insight from CRM data they collect--a process that requires the addition of fast and intuitive Web-based analytical reporting capabilities to existing CRM systems. First, let's define what nimble isn't when it comes to leveraging your organization's customer information assets:

waiting around for central IT to perform a symphonic CRM analytic overlay on top of your existing systems

a rip-and-replace approach that requires a CEO signature, seven figures, and months or even years to implement

taking specialized courses for data analysis or relying on trained professionals to interpret your results

getting bogged down in warehouse construction while real-time transactional data is flying by every minute of every day A nimble approach to reporting on and analyzing the demand chain involves adding strap-on tools to customer data, whether the data is in real-time transactional formats like XML; residing in existing sales, services, or marketing databases; or scattered in overgrown spreadsheets or "spreadmarts." Requirements for these analytic tools should include the ability to: Accept data from anywhere. The data warehouse might be a tab-delimited flat file taken from a spreadsheet or an SQL or Oracle database, or it could be constructed on the fly along with preformatted reports and user profiles from an XML transactional data stream. It's reasonable to expect easy handshakes between an array of demand chain data sources and your analytical reporting overlay. Adapt to existing application server and operating system infrastructures. It could be cubes from Microsoft Analysis Services, relational data from Linux-based systems, or a plug-n-play with existing Web servers. You shouldn't have to repay for IT infrastructure you already own. Deliver results in six weeks or less for less than six figures, using outside solution provider/consultants (just in case IT looks at you funny when you insist on the 6X6 project delivery). It can be done. Provide one-click interactive access to analytic reports. Customer intelligence systems too often take on the guise of Formula One race-cars, requiring a dedicated team of specialists to change tires, tweak the steering, etc. A nimble approach to analyzing demand chain data is more like filling your tank before the morning commute. It's a self-serve, no-fuss experience you can perform yourself without special training or a team of people to get you down the road to understanding customer reactions to and interactions with your organization You'll know your CRM application is properly integrated with the right analytics tools when a colleague comes down the hall and asks you for an answer to one of those previously painful questions that you both know is locked inside your demand chain data. Instead of politely switching the topic to the politics of the day, you'll be able to click on a URL, see the report that's of interest, then spontaneously analyze the data either in table or graphical format (or both), reducing millions of records down to the insight you need to drive the business. If your colleague wants to know more, you'll simply email the URL with the report you both created and she can dig further in the live data (it may have changed by the time she returned to her desk) to ensure you're making the best customer-driven decision for the business. About the Author Nathan Rudyk is vice president of marketing for Databeacon. Contact him at nrudyk@databeacon.com