I've written endlessly in this column that your CRM data is far more valuable (and expensive to maintain) than the system in which it runs. The \n\nHippocratic Oath for all CRM decision makers must be "above all, do no harm to the data." Even so, CRM systems must be maintained, \n\nextended, and integrated to meet competitive realities and business needs that evolve over time. And with each improvement to the CRM system \n\ncomes implementation cost, changes in maintenance\/operational costs, and risks to the data, so you have to consider the alternative of CRM \n\nsystem replacement. \nIs it time to rework, or replace? Here are ten real-world factors to consider about your CRM system on an annual basis, in priority order:\n1. Risk to Your DataSometimes, an existing system or its integration points cause data pollution: field corruption, record duplicates, or outright deletion. If this can't \n\nbe resolved due to architectural or operational reasons, I can think of no more compelling reason to replace a CRM system. 2. Users Hate ItSince human interaction is the source for most all interesting CRM data, unhappy users undermine both the functionality and credibility of the \n\nsystem. In evaluating the user-happiness issue, you have to look way beyond the sales rep's UI: look at every external touch point that's feeding \n\nthe CRM system. Do callers detest your IVR or your predictive dialer? Do customers get frustrated by your support web site? Does your \n\ne-commerce system have a ridiculous checkout sequence? If the central CRM system is an obstacle to fixing those problems, that's a pretty good \n\nargument for getting that system out of there.3. Too Hard to Integrate or ExtendMost enterprise software has a very long design life. Not so with CRM. Too many marketplace issues change the rules\u2014whether it's \n\nyour channel, your competitors, or the customer standards of service. So it's almost inevitable that even the perfect CRM system will need to be \n\nextended or integrated in fundamental ways every couple of years. Don't believe me? Just wait for the request for Twitter marketplace integration. \n\n If your current system is too hard to integrate or too expensive to extend, it's only a matter of time before you'll have to move on.4. Mobility and Extended AccessUsers in many organizations are increasingly decentralized and mobile. In highly virtual organizations, they may be consultants, contractors, or \n\npartner employees. If your system doesn't have a flexible security model, doesn't support mobile laptops well, and can't be used on an iPhone, \n\nyour field sales and support people are going to look increasingly weak vs the competition. This is not just a matter of flashy appearance. Mobile \n\nis the way people increasingly work. 5. Need to Replace an Adjacent SystemEven if your CRM system is in fine shape, you may have a corporate initiative requiring replacement of a system on which it depends. For \n\nexample, you may be rolling out a new multi-location call center that supports virtual workers. In making that replacement, analysts will note new \n\nrequirements for user features and technical integration points\u2014and your existing CRM system may not be able to support those \n\nextensions.6. DivestitureDivesting a company is usually very stressful to CRM systems because you need to partition the existing database, change all integration \n\npoints, redo the security model, and rework the user interfaces. In addition, there may be dramatic changes to the size and structure of the \n\nchannel, complexity of the price list, and the revenue-oriented business processes. These changes can be too much for an existing system to \n\nhandle. It may be cheaper and better to just start over.7. End of the Code LineAny time a system is a dead-end code line, whether it's been end-of-lifed by a vendor or it's an internal code base needing a full rewrite, \n\nyou're left with a three-way dilemma. Do you stay where you are, jump to the new code line (which is guaranteed to require new integrations, \n\nrebuilt custom extensions, database rework), or move to a totally new vendor? The choice will depend on risks, rewards, and time horizon. But \n\ngenerally speaking, you want to start planning to move to a new code line as soon as the old one's death is announced.8. No Web ServicesMore than any other category of enterprise application, CRM systems should be built on a solid platform of web services. "XML APIs" aren't \n\nenough: you need a service-oriented architecture to get the level \n\nof malleability and depth of integration required for a lasting CRM system. Business requirements, executive preferences, channel structure, and \n\ncompetitive pressure will mean you must modify and extend your CRM system over the years. So if your system doesn't have the ability to do full \n\ncall-in and call-out, using triggers, workflows, and asynchronous processing, you need to move to a platform that does.9. Cloud vs. On PremiseThe choice of SaaS vs. installed \n\nsoftware depends on the specifics of your company. There isn't one right answer for everyone. Here's an important truth: the vendors who excel \n\nat SaaS are not best-in-class in installed software (if they offer it at all), and the kings of installed CRM are really crummy at SaaS. So if your \n\nneeds have changed to the point that you really want to go for SaaS, or come out of the cloud for an on-premises solution, you will \n\nhave to replace your CRM system.10. Big New Feature FunctionalityNew software features are not good reasons to switch CRM systems. No matter how good the demo given to the VP of Sales might have \n\nbeen, it is rare for functionality to be the right reason to change. In CRM, the credibility and scope of the customer data is more important to your \n\ncompany's performance than any one piece of system functionality. Many of the things that look important and cool don't turn out to be important. \n\nIf a new feature set really is critical, other vendors will eventually deliver something that brings them to competitive parity. Of course, if there's \n\narchitectural functionality you need and your vendor doesn't even have it on their roadmap, your only real choices are to build a custom extension \n\nor buy from a new vendor who has what you need.David Taber is the author of the new Prentice Hall book, "Salesforce.com Secrets of Success" and is the CEO of SalesLogistix, a certified Salesforce.com consultancy focused on business process improvement through use of CRM \n\nsystems. SalesLogistix clients are in North America, Europe, Israel, and India, and David has over 25 years experience in high tech, including 10 \n\nyears at the VP level or above.