Tim Chou on True Technology Costs

Brian Wood Blog

CFO logoDr. Timothy Chou of cloud computing, Stanford, and Oracle fame wrote the opinion piece below for CFO.

AIS is proud to have forged a relationship with Tim over the past year; he is certainly a smart guy with keen insight.

(Note: you can see Tim keynote the March 6 CommNexus Headliner event on March 6 at Qualcomm entitled “Migrating to the Cloud”.)

Emphasis in red added by me.

Brian Wood, VP Marketing


The Truth About Technology Costs

When your IT people tell you that the new hardware and storage the business needs will cost a few million, remember: They can’t handle the truth. But CFOs can, and must.

The next time your IT staff comes to you with a purchase order for a server or storage says, “The price is $2.1 million,” do your best Jack Nicholson impersonation from “A Few Good Men” and growl, “Is that the truth? I don’t think so. You can’t handle the truth.”

The truth is the cost of that hardware is not $2.1 million. Oh, sure, that’s the one-time purchase price, but just like application software, it’s just the beginning of the cost.

Many years ago, when I was president of Oracle on Demand, we came up with a quick way for CFOs to estimate the cost to manage business applications. With the assistance of Gartner and IDC, we surveyed a number of customers and figured out that the cost to manage application software was about four times the purchase price of the software per year.

In fact, way back when, even Microsoft admitted that the initial purchase price of software is usually only 5% of the total cost of owning and maintaining a program.

Why? Because the human labor that will be necessary to manage the security, availability, performance, and change in the application (and all of the underlying software and hardware infrastructure) will quadruple the purchase price every year. Indeed, because the cost of application software continues to fall as the cost of labor rises on a global scale, my old rule of thumb may actually underestimate the true cost. And the same can be said for computers and storage.

Start with the cost of managing the security of the data center. I’m not just talking about the guard dogs out front (and the cost of Kibble n’ Bits). In order to make sure the building access system, the power and cooling systems, and the internal network are not full of viruses, you have to have the latest vendor-released software patches. That’s not a choice; it’s a necessity.

Then consider the software products you’ll have to purchase and the expertise you’ll need to retain to manage the availability of the network, not to mention the disaster recovery and redundancy expenses for your new servers and storage.

Of course, you’ll need new performance experts to do daily, monthly, or yearly capacity planning to insure the performance not only of your servers, but also, as you add boxes and expand your data center, experts to address the questions of power, cooling, and network bandwidth.

Finally, managing servers and storage would be easy if nothing ever changed. But everything always does. So you’ll need to factor in the cost of project managers, change management, and asset management that you’ll need to employ as the hardware you just bought becomes obsolete. And these costs will be minimal compared to what you’ll have to spend if you acquire a new organization and have to merge their data center with yours.

I think it’s safe to say that the cost to manage the compute and storage is at least two times the purchase price of the hardware per yearIn other words, that $2.1 million is not just the tip of the iceberg; it’s the tip of the tip. In just four short years, you’ll spend $16.8 million managing the security, availability, performance, and change in that server and storage.  If you don’t believe me, just look at how much headcount you carry managing the stuff you bought over the last 20 years.

So what should you do?  If you’re going to provide compute and storage cloud services to your internal customers, first make sure you’re not caught short and have allocated the budget to manage it. Second, if your compute and storage cloud service is specialized for your industry or geography, consider the possibility of turning that asset into a revenue producer by providing those services to others.  That will improve both the cost and the quality of the service as becoming a provider will necessitate both repetition and automation, and that will decrease your costs while increasing reliability.

There’s another option. Don’t write that $2.1 million check, and send your IT team out to look for compute and storage cloud service providers who can give you what your organization needs. Make sure they know the security, availability, performance, and change management features your company requires.  And ask the cloud service provider your IT people find to show you a few of the good men they’re going to use to provide the service. That’s part of due diligence; that’s the part where CFOs really deal with the truth. Only then will you know how to get both high quality and low cost in purchasing your next set of servers and storage.

Timothy Chou teaches cloud computing at Stanford University. He is the former president of Oracle on Demand and the author of Cloud: Seven Clear Business Models.