Dear Herb:

Here is the problem: We are a consolidated multi-company computer center. We bill our customers based on resource usage. The cost per unit of usage is based on our expenses to run the center. We project at the beginning of the year our estimated usage divided by our cost to run the center and create a cost per unit of usage. Customers are billed at the same rate throughout the year.

Our capacity planning group, after nine months, looked at the trend for 1997 and determined we are under-used, which means we will not recover all our costs. We are projected to spend $1 million more than our recovery rate. Our chief did not want to change the rate in the last 75 days of the year to recover our costs. So he decided to slash $1 million out of our expenses in the last 75 days of the year.

My question to you is should we shoot the capacity planners and if these guys are so smart, why are we in trouble?

Ed

Dear Ed:

Since I don't believe in shooting people, I suggest that you employ a volley of massive disdain instead. But I don't think the target should be the capacity planners, alone. If they were so smart, they wouldn't be in such a no-win job as trying to estimate computer usage. Take a hard look at how your boss is running the place.

Your cost-recovery model is indeed out of balance, but that can happen for two reasons. Either your capacity planners over-estimated the amount your data center would be used, or your financial analysts underestimated your costs. Regardless of which situation occurred, why did it take nine months to see it? I don't know what your chief was doing, but unless something very unexpected happened in the last quarter of the year, it certainly wasn't running the shop. Harry Truman had a sign on his desk saying, "The buck stops here."

In any case, your chief would be better off renegotiating his charge-out ar rangement. The type you have is almost always guaranteed to cause problems. Your contract ensures that if usage decreases, then costs either have to go down or prices have to increase. If prices go up, usage will decrease further as people find more efficient alternatives, causing prices to go up even more. The cycle repeats until the outsourcer takes over.

On the other hand, I wouldn't be surprised to learn that if you exceed your income, you have to rebate the "profit" to the users. If this is the case, as it is in many cost-allocation contracts, you wind up with no upside benefit, but lots of downside risk.

You are far better off with an arrangement where any surplus is rebated at the end of the year to the users and any under-usage is paid for by them in the same way. To smooth out the disruptions, quarterly adjustments can be done automatically, and everyone knows the rules of the game. At least with that mechanism, the needed capacity adjustments in the size or scope of your organization can be done in a logical and planned manner.

Go and talk to the boss. Maybe the capacity planners are not all that great, but unless your budget is humongous or your shop is unusually fat, ripping $1 million out of it in six weeks is guaranteed to crater morale and probably service as well.

You seem to have lost your place on the back page of InformationWeek. I hope that this is not permanent.

I think you missed a real opportunity to make Gornish, your CFO, squirm in "The Politics Of Outsourcing," but I think you still can recover.

Outsourcing is almost always viewed too narrowly by companies. The decision is never "whether we outsource" but "what do we outsource." We all outsource. We fly commercial airlines rather than own our own planes. We outsource everything from lawn care to the company cafeteria because we are not interested in being in those businesses.

So a way to push back at Gor nish is: "Thank you for raising the issue of outsourcing. As you know (directing your attention to Phil, your CEO) we already outsource many functions (company cafeteria, lawn care, travel planning, etc.) and I welcome the opportunity to expand this investigation to all of our major functions. I think any proposal that we take to the executive committee needs to address the strategic value of all of them.

For instance, it seems to me that many of the financial functions that we currently do in-house might better be "outsourced."

Dave P.

Dear Dave,

I am at least as sad as you are about having the column moved from the last page, but that is the way the magazine has decided to do things. Maybe someday InformationWeek will move it back.

So far as your approach which would counter our less-than-beloved CFO's parry with a suggestion to Gornish that he outsource some of his own activities, I don't think it woul d have the results that you and I would wish for to occur. Knowing Sid, he probably would simply say something like, "That may well be a good suggestion, and I pursue it, but stop being defensive and talk about why you haven't outsourced the systems in this company."

First let me say that I thoroughly enjoy your column. I really enjoyed the Outsourcing series (" The Politics Of Outsourcing " ; " To Outsource - Or Not ;" " No Outsourcing For Now ".

Here is my question: Do you have any samples of survey tools that you have used to survey your IT stakeholders? I'm especially interested in surveys designed to ascertain quality of service levels and customer-satisfaction levels, current and/or projected technology needs, or anything that would provide feedback on ways to improve the organization structure of the IT f unctional area?

Any help along these lines would be appreciated. Thanks in advance.

Greg K.

Dear Greg:

I am not a great believer in standard surveys because I think that the meaningful questions to ask vary heavily with the organizational culture and the business focus of the company. As an example, accuracy of reports (such as billing records) can well be of paramount concern in a consulting firm, whereas you might reasonably view online uptime as the big thing for an airline reservation system.

While I have a bias against cookbook questions, I do think it makes sense to follow a standard process in determining how you do a survey. This process should include how frequently to survey, the questions to ask, and finally, what mechanism to use to conduct the survey.

The first thing to do is decide the frequency of the survey. The frequency can be anywhere from after every customer-contact incident (the way man y automobile dealerships are measuring their service facilities) to something like a scheduled yearly survey. Personally, I opt for the "more frequent, the better" approach. I strongly believe that the time to capture information is when the event happens.

To determine the questions, many IT organizations make the mistake of asking themselves what things are important to users. It may never occur to them that those things important to the end-user are best determined--believe it or not--by asking said end-user.

Invite the key leaders out to lunch. Ask them what makes their businesses tick. Find out from them what they perceive to be the critical things an IT group has to provide to help them and help the company. Use their answers to build your survey.

Another valuable method is to use focus groups of employees who use the services day-to-day. Utilizing focus groups wisely is an art (see " But What's The Focus? "), so unless you have someone on your staff who is train ed in how to do it, get help either from your human resources department or from an outsider.

The mechanism of actually conducting the survey is vitally important and depends on the culture of your company. Whether you use comment cards, IT managers sitting down with important users, or a neutral third party will make a major impact on the usefulness of the results, so choose with care.

Keep in mind, no matter what you do, the key is to listen. It doesn't hurt to always ask, "How am I doing?" as former New York mayor Ed Koch did of practically everyone he met. And he got re-elected a lot of times.

I have been intrigued by the erratic behavior exhibited by businesses as of late. The decision to outsource or not is but an example of business worrying too much about the other guy. If your operations are better served outside your corporate structure, hire the outsiders. You can always evaluate the se rvice received and bounce them if performance is less than adequate.

Rick S.

Dear Rick:

Decisions to outsource some functions such as the cafeteria or the company's internal travel service are fairly easy to reverse. When it comes to activities such as IS, which provide the glue that hold together the modern corporation, I think we have a different story. I've rarely seen an attempt to re-internalize computer services go very easily--and I have talked to several people who have had to try. The major problem is that all of the knowledge as to how the systems run and how they all fit together, to say nothing of the personal relationships between users and the service provider, are in the hands of the outsourcer. Unless you physically acquire or re-acquire the staff that's providing the services, the task is a daunting one.

Having said all that, I still favor looking at outsourcing as a viable alternative to internal computer services. The different options available to a company under outsourcing agreements can provide a great deal of flexibility. Some people outsource running the data center, others outsource PC support and still others, the entire function. However, while outsourcing may make sense, it's a step that should not be taken lightly, because if a mistake is made, the divorce is not going to be a very easy one.