First Windows 7 Deployment Update: Under $26 Cost; $852 Annual Benefit

Rob Enderle

I got a number of questions both in this blog and in e-mail as a result of writing about the first big Windows 7 deployment. I thought it might be more interesting to address them in another post. The biggest question was how much does it cost to roll the product out, and one analyst posted a crazy number, so I figured I'd ask Baker Tilly. It turned out this was audited and reported in a study by Gartner.


In addition, I assumed since Dell was involved, it rolled out new hardware. It didn't, and this makes the energy and support savings more interesting. Finally, I left out a major partner that made this solution work more quickly, and that was Computer Associates. One of the reasons Del Monte hadn't deployed was because the Microsoft tools it wanted to use weren't ready yet, so Baker Tilly used CA Unicenter and was very pleased with the help CA provided.


Getting to $26 Per Desktop

Like many enterprise customers, Baker Tilly has Software Assurance, which means it has rights to a variety of desktop products covered by a renewable contract. Like most, this means it has pre-paid for Windows 7, and that eliminated the software charge. It also found that Windows 7 ran on 4-year-old Dell hardware, which is why Dell was involved, and the company believes it might be able to get five more years out of this hardware by upgrading to Windows 7. It didn't see a need for additional performance.


Baker Tilly used CA Unicenter, which it had also already purchased, to centrally manage the rollout and, using current exchange rates, this means it spent $56,000 for 2,200 PCs migrated to Windows 7. Most of this money was spent for tests and to get existing software to comply with the new operating system. This means its net additional cost worked out to $25.45 per PC over what it was already paying for services under the Software Assurance program.


$261 to $852 Savings Per Year Per Desktop

Baker Tilly stated it was saving around $161 a year in energy and services. It listed these as improved network access, automated deployment ($51) and management ($82), and PC power ($28) savings. This would work out to a nearly 700 percent return on investment, which might explain why so many companies are suddenly so interested in Windows 7. But this savings does not take into account the extended service life of hardware that it otherwise would have had to replace. That would be an additional $100 to $400 a year (depending on the cost of the hardware it would otherwise purchase) in capital cost savings.


In addition, given this was an aging XP base, there was a measured productivity improvement of two days for each worker (estimated at $591). Because Windows 7 (and Vista, actually) can be patched more easily in the background, this is the estimated aggregation of time workers would otherwise be waiting for their systems to be patched and rebooted. That's a total of $852 per desktop in savings as a result of an additional $26 investment.


Baker Tilly and Gartner provided these numbers.


Wrapping Up: With Windows 7 Your Results Could Vary

I can recall years ago when AT&T folks came over to a shop where I was working and told us that Windows 95 would save us $1,000 per desktop per year. Given we were only tracking $400 per desktop per year, I asked them if they planned to write us a check at year's end. Needless to say, we didn't use AT&T and while we did eventually move to Windows 95, we never saw $1,000 in savings.


Baker Tilly is a snapshot of what is possible in an environment where a key part of the rollout of the product was a locked-down desktop and solid central control. My expectation is that, were this control not in place, these savings would be substantially reduced. Still, even substantially reduced, these results are rather compelling, and it will be interesting to see if any additional companies can match them.


The other interesting thing is that while early adopter companies often get special support, they also often get the biggest number of problems. Baker Tilly indicated it hadn't seen any really major ones, which is very unusual for a company deploying this early. But it likely points to the fact that Windows 7 is a maintenance release and resides on top of and benefits from the years of Windows Vista experience. That means that this product probably is as good as Baker Tilly seems to think it is.

Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.


Add Comment      Leave a comment on this blog post
Oct 30, 2009 8:26 AM a. asdf a. asdf  says:

--"PC power ($28) savings."

I'm curious as to how they actually determined power savings. Did they hook up a power meter to a XP system and an identical Win7 system for a week? And where exactly is Win7 getting these additional savings from? The only real details I've found on this topic is at http://blogs.msdn.com/e7/archive/2009/01/06/windows-7-energy-efficiency.aspx They talk about slowing the processor clock, but that is assuming your processor supports it (things like Intel speedstep, which wasn't available on desktop CPU's a while back). XP also had the ability to scale down the processor clock.

--"It also found that Windows 7 ran on 4-year-old Dell hardware"

That is certainly possible, but again, as far as I know, the desktop CPU's 4 years didn't support lowering processor clock. 

And what about losing memory to the OS? Win 7 requires 1GB vs 128 MB XP. And hard drive space? Win7 (16GB) vs XP (1.5GB).

--"Because Windows 7 (and Vista, actually) can be patched more easily in the background, "

I don't see why you can't just set up XP to update over the lunch break or after 5pm on Tuesdays.

Oct 30, 2009 9:37 AM Rob Enderle Rob Enderle  says: in response to a. asdf

You're thinking too much.  The power savings is largely coming from turning the machines off or suspending them when not in use.  Something they weren't evidently able to control under XP centrally as well.  You've been able to lower the clock speed of the processor for some time, this is an automatic function for thermal management in place since the mid-90s.  Typically we try to go the other way, its called overclocking.   Current hardware typically allows you to do this more dynamically rather than through settings though.

We've had storage headroom for much of this decade.  This is a financial company and would rather not have a lot of storage on PCs anway, the loss in capacity might actually be considered a plus because they would likely rather have the information remian on secure servers.  

In a lot of companies people work throug lunch or take staggerred lunch breaks and work after 5.   If the system is off you need to be able to wake it up to patch it.  I can recall watching people actually have a patch reboot thier XP systems the middle of a powerpoint presentation.  Auditors tend to work long hours and often don't take set lunches.   So you could but it would disrupt the work. 


Oct 30, 2009 7:06 PM Ken Hardin Ken Hardin  says:

Hi, Rob -- Ken here. Just wondering, is it typical for large shops to simply not factor in some swag at the proportional cost of a software upgrade beneath a program like SA if they are taking as hard a look at a rollout as Baker Tilly obviously is? Obviously, they'd be under SA anyway, but it seems that some cost-recognition would be in order, if for no other reason than to justify the overall SA outlay. But that's a cost center manager from an SMB talking, and of course we try to count everything. Thanks, Hardin

Oct 30, 2009 7:46 PM Rob Enderle Rob Enderle  says: in response to Ken Hardin

Typically for a roll out of a product you look at the incremental extra cost against measured benefit.   The strongest benefit here is likely the ability to extend the life of the hardware and reduce ongoing support costs for the result.  

But when SA comes up for renewal the cost saved would be factored in to justify the SA renewal.  Remember that SA cuts across a variety of products (mostly server side) so the estimated $100 per desktop software cost savings (at volume rates) would, in addition to the savings for their coming server and exchange deployments (which are in plan) justify the SA renewal cost. 

Now it might be interesting, after the fact, to roll up all the costs and take them against the alternatives but this is rarely done.   IT shops typically operate project to project and don't have the bandwidth to do analysis that cuts across a series of projects (often this is because each project is generally staffed differently).

In the end, and what we often forget, is the true benefit of SA isn't the savings but the ability to avoid budget surprises by making spending predictable.   Generally it is worse to blow out a budget than it is to overspend because budget surprised reflect on management competance and overspending isn't as visible. 

This is one of the things I'm not sure folks like Google or the Open Source community really understand. 


Post a comment





(Maximum characters: 1200). You have 1200 characters left.




Subscribe Daily Edge Newsletters

Sign up now and get the best business technology insights direct to your inbox.

Subscribe Daily Edge Newsletters

Sign up now and get the best business technology insights direct to your inbox.