A lot of Walmart's success is due to management's early and continuing involvement with IT and their consequent support for the early adoption of what later proved to be outstandingly successful technologies including System 38 based retail management applications, information warehousing, and supply chain management.
The number one strategic lesson IT management should take from Walmart, however, has nothing to do with any of these: it's the simple recognition that when your product is equally available from other people, you have to compete first on price and second on your service image.
Notice that I say "service image", not service. Walmart's greeter is an image expense, humanizing the store and thereby simultaneously reducing the resentment incentive to shop lifting and giving customers an emotional basis for the perception that service is better than it is.
The lesson for IT here is simple too: if your IT architecture (hardware, software, and people) is the same as everyone else's, then you get ahead the same way Walmart does: by combining cost cutting with service image improvement.
Cost cutting has strategic consequences: because everybody uses the same software on the same hardware with the same interchangeable people, any cuts you make, they make - and nobody's further ahead afterwards. Cost cutting, in other words, is a game at which you can lose but not win.
If you've currently got an out-sourced PC help desk you have a one shot opportunity for improvement: bringing your greeters back in-house will improve user service perceptions even if the actual service doesn't change at all - and you'll get your additional costs back through reductions in resentment inspired user actions like carelessness with Wintel media, data theft, and hooking up their own devices.
But beyond that, what? Cost cutting is a downhill street with a cliff at the end - and , like Walmart, you can improve your service image but you can't afford to actually improve service much. So what do you do?
My answer is that the only long term solution is to get off the treadmill; dump the source of most of your costs and risks: Microsoft's client-server architecture - put in something that's simpler, more reliable, and safer: meaning Unix, smart displays, lots of open source applications, and putting your money into empowered IT staff working directly with and for users.
In practice that means starting with a skunk works, hiring some new people, and exploding slowly -because an immediate technology about face isn't just risky, it's usually suicidal. But, remember, if you don't get started, your successors will still be fighting today's battles next year, and the year after, and the year after that - just like you're doing this year, and did last year, and the year before that.
There may be a different answer, but I don't know of one - and no matter how you look at it the PC business is now at the stage where the stuff we tend to focus on: hardware, software, licensing, and infrastructure, constitutes such a trivial portion of overall IT costs that there's really nowhere left to go - except productivity driven architectural change, organizational change, and the turning inside out of IT control structures to directly serve the user.