I was taken by a customer presentation by Stephen Rayda, CTO of Purdue Pharma, at this week's EMC Analyst Summit. Like all pharmaceutical firms, Purdue faces a lot of pressure \u2013 and it just got a new CEO to boot. He asked which departments didn't need his attention; luckily, IT was the only one to make that list.\nThis is in sharp contrast to what typically happens. When a new CEO arrives, IT becomes such a priority that the CIO polishes up his or her resume, anticipating a catastrophic career change driven by the new CEO. CIOs have a relatively short lifespan. They effectively start out as an endangered species. We don't focus on that problem as much as we should.\nHow Purdue Pharma Fixed an IT Infrastructure Mess\nAt Purdue, like many companies, IT started to get out of control \u2013 a mess of mismatched hardware and software, all heroically (and often unsuccessfully) trying to interoperate while the vendors that sourced these technologies pointed to each other as the core problem to be solved. Meanwhile, it should be impossible to fix a mess such as this simply by patching, but that's exactly what most CIOs attempt. Rather than fix the core problem of excess complexity, they just make it worse. That's why IT typically tops any list of departments facing critical problems.\n[ News: Cisco Said to Be Selling Most of VCE Stake to EMC ]\nRayda saw this and prescribed VCE's Vblocks. These replaced the mess with a known configuration of things designed to work together. He reported cost savings, energy savings, increased customer satisfaction and, best of all, being removed from the problem list and placed at the top of the "Don't Break What's Working" list.\u00a0\u00a0\n(Note: The author is assigned to both VCE and EMC through his work with Valley View Ventures.)\nAs part of the process, VCE took over most operational responsibilities. As Purdue moves to higher levels of automation, the staging decisions now happen inside VCE. This removes the conflicts of interest inside Purdue, allowing it to just focus on return on investment for any related efforts. People willing to embrace the change were retrained to fill user-facing roles focused more on expanding services and less on keeping stuff functioning. Those who didn't want to change were managed out of the company.\nOne interesting benefit was that, when Oracle came in asking for an additional $1 million in licensing charges, based on new core loads, Purdue was able to shift loads, reducing that unplanned charge by 90 percent thanks to the flexibility of the system. Typically, that doesn't work so well.\n[ Commentary: How VCE Created an Amazing Joint Tech Venture ]\nVCE managed to remove the complexity and, indirectly, the folks who didn't want to change. This let IT shift focus from just keeping things running to providing real value. The reported benefits included a 75 percent reduction in data center footprint, a $9.5 million reduction in annual costs, $2.5 million in annual power savings, and far happier customers. Rayda, meanwhile, went from being on the executive endangered list to the only top executive who the CEO doesn't want to muck around with.\u00a0\u00a0\nFor Better CIO Tenure, Pick the Right Vendors\nThat last point got me thinking about how we measure IT vendors. In this new world of analytics, wouldn't it be interesting to rank vendors by the stability of the CIOs who favor them? You'd need to create a database that's frequently updated, as the average tenure for today's CIO is pretty short, and rank vendors by the average tenure of the executives who use them heavily.\nThis would give you a sense of the vendors, in aggregate, that take care of their customers and, as a result, ensure a CIO's longevity instead of providing an unplanned, undesired premature retirement outcome. Promises and technology aside, shouldn't we choose vendors that assure and protect their customers over those that don't?\u00a0\u00a0\nLately I've been looking at net promoter scores, now in wide use by a number of vendors, to see who takes good care of their customers and who just feeds off them. While NPS is important, I think the tenure and success of the CIO may be an even better indicator of which vendors are healthy, at least when it comes to addressing the problems that most face.\n[ More: 4 Tips for Choosing an IT Vendor and 6 More Tips for Choosing an IT Vendor ]\nI'm reminded of a story my grandmother told. Once upon a time, a woman put salt in her tea instead of sugar and then went door to door asking for advice on how to fix it. Each person suggested a different herb or ingredient that would offset the salt. These fixes only made it worse. Finally the woman reached the smartest woman in the town. She simply said, "Start over." Most IT shops resemble the bad tea, with layer after layer of complexity on top of a problem that has been there since the beginning. The best (and cheapest) course of action is to just start over.\nI volunteer on a Jaguar Forum, and I can't tell you how often we provide similar advice: Rather than tell car owners to try to fix a series of ever more complex problems, we recommend they just sell the car and get one that's in better shape. We used to call this a forklift upgrade when I worked in telecom.\nThere's a big difference between vendors that contribute to a CIO's health and longevity in the job and those that contribute to CIO turnover. Some may argue that CIO turnover can lead to upward mobility. However, there's always the risk that you'll get the job \u2013 suggesting that, in all cases, picking vendors that take care of their customers over those that contribute to their premature retirement might be prudent.