- The Design Team
- Purpose, Objectives, and Scope
- Cost Justification: Return on Investment and Total Cost of Ownership
- Recommendations on Areas of Significant Cost Reduction
- What's Next?
Cost Justification: Return on Investment and Total Cost of Ownership
Like any project of this magnitude, the Windows 2000 migration requires the completion of a cost justification. It is important to include in the team, at least at the beginning, someone capable of making such an analysis. You undoubtedly will find that there are increased costs with the Windows 2000 migration because of significantly greater hardware requirements and the purchase of Windows 2000 software, licensing, and any application software upgrades required for compatibility with Windows 2000. Other costs include the time spent by the design team, application certification (a potentially a big expense), and actual migration costs, such as debugging installation procedures during the pilot, downtime while users are migrated, and expense incurred in network upgrades, if any.
Chapter 1 suggests that this is a good time to clean out all that old legacy hardware and to leverage the migration as a way of obtaining new hardware. Unfortunately, in many situations the purchase of new hardware is restricted or not feasible at all. Many companies are trying to keep their legacy hardware486s especiallyto stay within budget restraints. While Windows 2000 requires Pentium-grade processors and significantly more memory than NT 4.0, there are ways to get a little more life out of that old hardware. In fact, Terminal Server is included in the Windows 2000 operating system for the express purpose of allowing 486 machines to participate in the Windows 2000 domain as Terminal Server clients running applications. This is an attractive way to get the most out of budget funds.
Cost justification of Windows 2000 in the late stages of the beta was a much debated topic. In September 1999, The Gartner Group issued a report claiming that the migration cost of moving to Windows 2000 Professional on the desktop could take more than three years to realize a return on investment. The Gartner report claims that it cost between $1,250 and $2,050 per desktop to migrate from Windows NT Workstation 4.0 to Windows 2000 Professional, with the migration cost from Windows 9x to Windows 2000 being between $2,015 and $3,100 per machine. The report also claims that customers will not see a return on investment (ROI) for more than years. These figures apparently included only migration of the desktop, not the infrastructure. According to Microsoft's response, Gartner did not take into consideration installation tools, such as Remote Installation Service and unattended setup.
Unfortunately, I was never able to find the actual report, even from Gartner's Web site at http://www4.gartner.com/Init. Microsoft's Web site has an article that refers to it and, of course, discredits it. However, Microsoft does reveal the figures Gartner came up with. The Microsoft article, "Gartner Group's Windows 2000 Professional Migration Model," is located at http://www.microsoft.com/windows2000/news/bulletins/gartner.asp.
The Arthur Andersen Report
As expected, Microsoft was not happy with the Gartner report: The company declared that the model was flawed, and responded to it on the Web site http://www.microsoft.com/windows2000/guide/server/reviews/roi.asp by quoting a report of its own, produced by Arthur Andersen.
By its own admission, the Andersen report, "Microsoft Windows 2000 Server: ROI Impacts for Corporate Customers," "makes generic high-level ROI observations about the impact of Windows 2000 Server...as a starting point in assessing the financial impact [of] Windows 2000 Server." The report does not use a case study to calculate the ROI to produce specific numbers like the Gartner report does. It simply makes observations without any hard data (none is presented, anyway). Still, the report contains excellent information and could be a good resource in your effort to get a handle on ROI calculation. This report compares Windows 2000 to only Windows NT 4.0, Service Pack 5; like the Gartner report, it analyzes only desktop migration, not infrastructure costs.
Because this report is a 107-page document, I have summarized some important points in the following sections.
Andersen used Microsoft's Rapid Economic Justification (REJ) Framework, available from the Microsoft Web site, as a framework for evaluating a potential ROI. Although the report notes that Andersen does not endorse this tool, it is impressive that they choose it. The REJ framework is very different than other ROI tools I found because it defines a framework for evaluating Windows 2000; it's not a number-cruncher. It definitely deserves evaluation for your ROI calculation.
The REJ tool and documentation are available at http://www.microsoft.com/business/default.asp.
The test scenarios Andersen used for this study are interesting because they identify the areas in which Andersen felt the greatest savings could be found:
Scalability and performance
Interoperability and security
Key Findings of the Andersen Report
The Andersen report made the following recommendations as a result of testing. Remember that this is a comparison of Windows 2000 to Windows NT 4.0 SP5:
Windows 2000 is more scalable, reliable, and manageable.
There is a significant learning curve in implementing Windows 2000. Use this against cost savings.
Windows 2000 will have a significant positive impact on overall cost and control of IT assets.
Windows 2000 features will have different value to different companies, based on size, complexity and business mission.
Windows 2000 has sufficient benefits that it merits evaluation to see whether it will give value to your company.
Benefits of Windows 2000 are fully realized only when Windows 2000 Professional is implemented.
The greatest value of a Windows 2000 implementation is realized when integrating features with the operating system, resulting in improved reliability and scalability. It allows close integration of client and server.
Larger organizations are likely to see a shorter ROI period than smaller businesses. Savings in IT management and administration alone might justify Windows 2000 migration for a large company. In addition, large organizations will more easily overcome training costs than small ones. This is not to infer that small organizations can't cost-justify the upgrade; just that the payback period will take longer.
Windows 2000's Most Critical Features
The Andersen report identified the following features as most critical to the organization, assuming implementation of Server and Professional products. The report also explains the benefits that are not detailed here. This is simply an overview to demonstrate what this report determined to be the most important features in calculating the ROI:
Enhanced Distributed Services with Active Directory standards-based security, and centralized management tools.
Enhanced reliability and availability via Dfs, Enhanced Clustering Services, and reduction in reboots.
Enhanced scalability via symmetric multiprocessing, network load balancing, and Extended Memory Architecture.
Enhanced application support via COM+.
Enhanced interoperability (via X.500, LDAP).
Summary of ROI Reports
Neither the Gartner nor the Andersen report gives a template or a set of calculations that can be done to produce an ROI analysis for any given situation. The Gartner report points out some very disturbing figures in terms of desktop migration cost. The Andersen report, on the other hand, finds favorable reasons to migrate to Windows 2000 Professional and gives a list of ways to gain a favorable ROI, even though the report doesn't mention specific numbers.
The point here is to expose the designer to two differing opinions and calculation methods, and then allow him or her to examine the issues and learn how they might apply to a particular enterprise anticipating a migration. The designer then can identify any stumbling blocks in advance of the migration.
The Gartner Web site, at http://gartner3.gartnerweb.com/public/static/win2000/win2000.html, includes a number of excellent articles on Windows 2000 migration. There are three reports of particular note: "The Higher Licensing Cost of Upgrading to Windows 2000," "Windows 2000 in 2000: Cases for and Against Desktop Migration," and "Windows 2000 TCO and Migration." This third paper is particularly useful because it addresses ROI, migration costs, the process of creating a timeline, and Windows 2000 project teams (referred to in this chapter as the design team), and it provides some interesting numbers on staffing requirements for the migration. Unfortunately, it has more pretty pictures and icons than hard data, but it does have some good information and some useful data.
The Problem with IT ROI Calculation
The problem in calculating information technology (IT) ROI was described in the article "What's Your ROI," in the August 24, 1998, issue of Information Week Online (http://www.informationweek.com/697/97iuroi.htm). Author Jeff Sweat notes that although there is more pressure on IT managers to justify IT projects with a calculated ROI, few have figured out how to do it. He also asserts that the many tools built to automatically calculate ROI are unreliable. He further claims that the reason for this difficulty is that IT affects so many aspects of the business that it is difficult to tie IT upgrades to a cost reduction or productivity improvement.
Establishing a hard cash benefit of something that hasn't been implemented is difficult and, in some cases, suspect. For example, trying to predict that implementation of Active Directory will save money because advanced power management is more efficient (as one presenter tried to prove), or because people can do their job faster, or because administration will be easier, is difficult. You have to be able to tie those results to actual costs and show that the new technology can reduce costs. You might show that administrators can handle more tasks and reduce the number of administrators, or that this will allow business expansion for some time without hiring any more administrators. Logon time to a catalog phone order specialist is critical because they are evaluated on the number of calls taken, and slow logon time can reduce the time they have to take orders.
At a Windows 2000 conference I attended, I witnessed a presentation on cost justification methods. The presenter had a slide claiming that improved power management would keep laptop batteries charged longer. He claimed that if it saved one minute a day, this could be extrapolated to a savings of $1.5 million, by multiplying that minute by the number of users with laptops, and by the number of days in the year. The attendees quickly pointed out that this is an immeasurable saving. This is an example of what I refer to as "soft money"it doesn't really translate into savings that you could ever measure or validate. Besides, if you get into a contest of saving seconds for various operations, you might lose a few battles on tasks in which Windows 2000 might not be as efficient as another OS.
ROI Calculator Tools
I explored a few ROI calculators available on the Internet: All tend to be very specific to an industry or product, and they use numbers from surveys. For instance, if you want the ROI on implementing Lotus Notes in an organization of 10,000 users, a tool would find similar companies and see what they saved (via a survey), average them, and return information (assuming that you will be average).
Every company is different, though, and your ROI calculations will be specific to your business. It requires some time and effort to get an accurate ROI. Don't trust the success of the migration to a five-minute ROI calculation (as one vendor claims).
To calculate a believable ROI, you must have hard numbers. It is easier to not only accurately predict the ROI of the Windows 2000 upgrade, but also to gain support of upper management when you have a specific project that would benefit from the upgrade. A discussion on how to identify those hard costs appears later in this chapter in the section "Identifying 'Hard' Cost Savings."
You should keep two important points in mind when creating a valid ROI analysis. First, management typically isn't interested in all the technical goodies that we computer-types get excited about. So what if you have a hierarchical namespace? So what if Windows 2000 uses industry standards like x.500, LDAP, and DNS? So what if this is closer to an enterprise operating system than NT ever had hopes of being? So what if administration is much more granular than NT? Management wants to know how many people it can lay off, how much more work can be done, how production of a product can be increased, how the number of computers can be reduced, or how other overhead can be reduced. Secondly, if you can find a particular area that you know will have an immediate benefit, use it and see how far it takes you. This should be a visible, high-profile project whose success is all but guaranteed. The scenario cited earlier in which Compaq reduced the number of servers is a good example of this.
The same approach should be taken when cost-justifying a migration to Windows 2000. Study and test the benefits of Windows 2000, and review with the design team the area or areas of the company that could benefit most from the upgrade. Chapter 1 discusses a number of features and benefits and shows how they can be translated into cost savings or justification of the upgrade. As more companies adopt Windows 2000, more accurate data from other sources will become available.
This section lists areas associated with implementation of Windows 2000 that affect implementation costs. Rather than try to list individual costs, I have listed areas that likely require additional funding. As the designer, you must rely on the design team to try to predict actual costs so that no surprises cause the project to run over budget. These are only representative costs, and they may be more involved for your situation. Still, this should give you a good start. Following are the areas that affect implementation costs:
Hardware: Workstations (new purchase and upgrade).
Hardware: Servers (new purchase and upgrade).
Training (end users).
Increase in capacity (new subnets, routes, wiring, and switches).
Additional WAN costs (more service coverage, additional lines, and line upgrades).
Downtime during testing.
Increased bandwidth consumption during pilot.
Interoperability. During the migration, when there is a mixture of Windows 2000 and NT Servers (and perhaps MAC, UNIX, and NetWare clients) and when users are trying to connect to resources, the potential for downtime is significant because resources may become unavailable.
Purchase and licensing of new applications to replace legacy applications that are not Windows 2000-compliant.
Modification of custom (in-house developed) applications to make them Windows 2000-compliant.
Testing for Windows 2000 compatibility.
Licensing of Windows 2000. See Microsoft's Web site, located at http://www.microsoft.com/windows2000/guide/server/pricing/default.asp, for pricing information.
Increased administrative costs.
Security (more complexmore time to design, configure, administer).
Ramp-up. In the long run, administrative costs will be reduced, but initially the administrators must get up to speed, and it will take time to get the process tuned.
Time and salaries of the design team and other staff. What will it cost you to fund this team?
Tool development, especially migration tools, custom administrative scripts, and so on.
Increased number of users. Will more users be added than were in the previous environment? This could happen if a consolidation occurred, such as converting a group that was using NetWare to Windows 2000 in an effort to move to a single OS platform, thus increasing administration costs of supporting these new users. However, it might be possible to train the NetWare administrators and use existing resources to some degree.
Hand in hand with ROI calculation is the total cost of ownership calculation. An excellent description of TCO calculation is found in a report by the Gartner Group. The article, "TCO Best Practices Amplify IT Cost Reductions,"(http://www4.gartner.com/Init) defines three areas of best practices in calculating TCO: technology, process, and people. The following lists categorize these:
Technology best practices:
Automated asset management: systems management and scalable architecture
Fault tolerance: software inventory and automated software distribution
Process best practices:
Managed user environment
Service-level tracking and management
People best practices:
Stable IS organization
Especially intriguing about the report's findings are the graphs illustrating the different TCOs if all these practices are followed, versus the TCO if none or any two best practices are followed. Of course, the lowest TCO is realized, according to the report, if all the best practices are implemented. However, the factors contributing to the savings, in order of impact (highest impact listed first), are these:
End user IS
Gartner concluded the study with the following comment:
The deployment of technology carries with it both costs and value; poor implementation occurs when the costs outweigh the benefits, producing negative value. Technology itself is seldom to blame; most often the cause of negative value is a poorly scoped implementation or the failure of the organization to integrate technology into the process. Clearly, the correct combinations of technology, process, and people-related best practices have a substantial impact on lowering TCO.
This statement should serve as a warning to anyone planning a migration to Windows 2000. Just because you have planned it out on paper and have a positive bottom line financially doesn't necessarily ensure success or a cost analysis at the conclusion of the project. You must implement a realistically scoped plan properly, and employ technology-, process-, and people-related best practices.