Randy Mott is out as Hewlett-Packard CIO after a shake-up by new CEO Leo Apotheker. Among Mott’s many accomplishments over his more than 30-year career at Wal-Mart, Dell, and HP, he led a massive, three-year consolidation and centralization of HP's IT, cutting costs, staff, data centers, and applications, as well as the average time it takes to finish an IT project.

While Mott made plenty of waves at HP while driving that transformation, he is also among the most respected CIOs in the business. (Mott was InformationWeek's Chief of the Year when he was CIO at Wal-Mart, and he's on our editorial advisory board.) In numerous interviews with InformationWeek dating back to 1996, when Mott was CIO at Wal-Mart, he has shared with us his high-level but always practical thinking on IT strategy. Here are some of the insights from those past interviews:

1. Know The "Revenue Of IT"

If a company is looking only at how much IT costs, that figure will always look too high. So a cornerstone of Mott's philosophy is to put a measurable value on the work the IT organization does--what he calls the "revenue of IT." That data represents all the benefits, both hard dollar and intangibles, that a project delivers in the 12 months following full implementation.

Said Mott in 2008: "Every business has revenue, but IT typically doesn't ... because we don't have the discipline to capture the benefit of projects in a way that we can show the CEO or executive committee and have numbers that are real."

That's part of why Mott drove hard at HP for a cost-benefit analysis for every IT project that was agreed to by business unit leaders and their finance teams. Business leaders sometimes resisted, leading to what HP IT staffers have described as "don't blink" moments. But the cost-benefit analysis gave the "revenue of IT" figures credibility with execs because they weren't just IT's numbers.

2. "It's Tough To Imagine Fast Enough"

This idea is about getting technology into people's hands quickly, and not waiting to have every detail and use case figured out. It's about letting people on the front lines experiment with technologies before all the technical problems are worked out. To really understand a technology's potential, IT needs to put it in end users' hands. Said Mott in 2003: "It's tough to imagine fast enough. You have to experience it to imagine what's possible."

3. Fewer Projects At Once, Finish Them Faster

Mott pushed HP's IT organization to do fewer projects at a time but to finish them faster. He said HP got its average project time down to six months in part by putting more people on a given project while shortening the deadline. The goal isn't fewer projects in total; it's faster turnover. This year, Mott had set a goal to squeeze that average time further, to 90 days.

This ties to the previous point, the need to get technology into people's hands sooner so they can experiment. IT can't set priorities for 10 projects spread over the next two years, he said, because once projects one, two, and three are done, the results may change what would have been four, five, and six. "There's an innovation feedback loop that right now, in these long development cycles, IT constrains," Mott said in 2010.