Gartner calls managing customer-related data the number one obstacle to getting a return on CRM investments. To fully exploit the power of CRM, and to realise its much-hyped promise, organisations must first recognise that customer-related data often resides in dozens, sometimes hundreds of separate databases, files and data feeds; they must then understand how to utilise all these information sources in a consistent way. Complicating this integration is the thorny fact that frequently business definitions differ from system to system: what is a "customer" in one system may be a "household" in another. Or, worse, what makes up the customer record in one system may be completely different -- or even conflict with -- what constitutes a customer record in another system. Despite spending millions on CRM systems for call center scripting, campaign management software and the like, companies can't accommodate a marketing department's simple request to link direct mailing or call center results to Web site hits for a given campaign. Thus, data integration can't happen without consistent data and comprehensive customer-related business definitions across the CRM infrastructure -- e-mail, Web system, call center, direct mail, data mart, campaign management and so on. Most organisations have sharply delineated business units around products versus the customer, resulting in wide rifts between sales, service, marketing and product lifecycle management. Why, for instance, should channel sales share leads with direct sales? Information technology (IT) departments bear the brunt of this departmentalization. Corporate executives, moreover, are frustrated by their inability to use the data contained in separate business units, which has been gathered over the years as a result of concerted marketing and CRM initiatives. The high cost of these siloed CRM projects only adds to the pain. The companies struggling to make CRM work today originally focused on solving specific operational problems, at which point CRM systems did very well. Unfortunately, in the process they left widening gaps between their data pools and CRM systems, as business units added new CRM modules and services to the mix. The various systems that make up the middle layer of CRM -- the operational side, such as sales force automation and call center personalisation systems, and the analytic applications such as data mining and targeted marketing campaigns, for example -- work very well inside their boxes. But nothing has coordinated the customer touch points, or bridged the gaps between the siloed systems that support them. These problems are more painfully apparent now that the two-year vacation from capitalism is over. Companies need to maximise the huge investments they've made and concentrate on getting things to work the way they wished they would have in the first place. Meanwhile, operational efficiency is no longer the main goal of CRM; further, marketing has become a primary CRM driver.