Evaluating the Costs of Multivendor PC Hardware

Although some level of diversity in hardware suppliers is typically unavoidable and beneficial, it does introduce costs that must be understood during the vendor selection process. This is particularly true for PCs and other end-user devices. Organizations should strive to keep the number of active vendors low.

META Trend: During 2003/04, client standardization and managed build/distribution processes will enable an adaptive and cost-effective end-user computing environment, with focus placed on certification rather than technology homogeneity. Through 2005/06, IT staff will face challenges in managing pervasive client devices. By 2007/08, client computing models centered on IT group device ownership will yield to managed subscription services across corporate and personal devices.

The ongoing focus on cost reduction in the end-user space has led to procurement practices (e.g., reverse auctions), which in turn are introducing multiple PC vendors into corporate environments. However, restricting the number of vendors within an IT environment is an important factor in controlling both upfront and operational costs. Organizations need to carefully examine the impacts vendor diversity will have on their overall cost of ownership. Although increasing commoditization has reduced the functional differences between vendor product lines, vendor processes and detailed product-support issues remain. How much impact the mixed-vendor environment will have on procurement costs, operational costs, and service levels will be dependent on numerous factors within the organization such as levels of outsourcing, management processes, and environmental complexity.

Having multiple PC vendors can be beneficial by reducing dependence on a single vendor’s ability to supply product. This can be critical in reducing exposures in large-scale time-sensitive rollouts. Currently, specific product shortages generally tend to be relatively rare and short-lived. In addition, when shortages do occur, the causes are frequently cross-vendor and impact several players at the same time.

The increased costs of the multivendor environment are broken into two groups: those resulting from having an additional vendor and those resulting from having to support additional product lines. On the vendor side, many of the costs will relate to procurement processes. Most organizations have well-established processes and systems for procuring computer equipment that are increasingly integrated with vendor systems (e.g., Dell’s Premier Pages). Each vendor will typically require variations to existing processes or even the development of new processes/systems. Procurement issues will weigh more heavily if different channels (e.g., indirect reseller channel versus direct vendor model) are involved.

In addition, splitting a single large bid lowers the overall value to any given vendor and diminishes the chances of getting a strong ongoing discount. The traditional belief that multiple PC vendors can also be positioned against each other to drive lower costs is giving way in the current competitive market to specially priced exclusive deals. Many vendors will provide additional discounts of up to 10% for exclusivity.

Furthermore, strong business relationships are hard to develop if too many vendors are included. For at least the first six months of any new partnership, care must be taken when scheduling time-sensitive business-critical projects. This will create overhead in the form of additional oversight, as well as more frequent status meetings. In addition, each negative event will require some level of follow-up ranging from a deskside visit to a formal postmortem review. It is not unusual to see up to 5% of system purchases have some form of error (e.g., missing cable, wrong image loaded, misdirected shipment) during the first six months of a relationship.

Product-related costs can be even more substantial. Although PCs and servers are frequently thought of as interchangeable commodity devices, there remain significant implementation differences between vendors’ product families. These differences manifest themselves in several areas, including the following:

• Deployment/implementation costs: Deployment and implementation costs will increase as technical staff will need to deal with variations in the implementation processes. Particularly when large numbers of systems are involved, the ability to simply replicate a solution with minimal variations reduces implementation costs (each system variation will add 25%-50% to image preparation costs) and improves problem resolution times (each variation can increase problem resolution times by 10%). Introducing a new hardware vendor can delay an implementation cycle by 6-10 weeks, as IT staff develop and test a new working image for the new platform.
• Option selection: The selection of options can be impacted because each vendor will have different add-on components. This limits the ability to mix and match parts between systems and necessitates the stocking of additional spares. This alone can increase support costs or slow problem resolution. Even if a service provider handles this, it can significantly add to problem resolution in remote locations and will inevitably add to the costs of a support contract.
• Service and support agreements: Service and support agreements will frequently become more expensive when multiple vendors’ product lines must be covered. On the desktop, this typically adds up to $2-$5 per month per desktop. On servers, the increase can be 7%-10% on the contract. In some cases, service organizations have been unwilling or unable to properly support another vendor’s products. In one case, a customer using one vendor’s service organization was unable to get satisfactory support for a second vendor’s systems. The customer eventually dropped the second vendor even though its systems were significantly better priced because the service contract had a higher priority.
• Training costs: Training for technical staff will increase. On the server side, for every new platform, each technician will likely require at least one to two days of classroom training to become familiar with new tools and configuration options, while senior technicians may require more in-depth one- to two-week-long courses, particularly if the organization is self-maintaining. For client systems, the impact will be less significant. At a minimum, organizations should expect to spend at least double what is currently spent on vendor-specific training if an additional vendor is added to maintain equivalent coverage. This can increase even more if higher levels of cross-training are required for geographic dispersion, vacations, staff advancement, etc.
• Manageability: Organizations that have invested heavily in leveraging the capabilities of a vendor’s management tools will find that some functionality will either be lost or need reworking for other vendors’ products. Running parallel management environments creates the potential for confusion and a need to retrain support staff. Although different tools can frequently roll up into a single combined console, the functionality and actual interaction with the tools remain distinct. Support personnel will need to be aware of and account for disparities and will find that management processes (e.g., scripts) cannot be directly replicated without significant work. This will add to ongoing maintenance, problem detection, and resolution activities. It would be reasonable to expect a 10%-15% hit on those times and, in a highly diverse environment, as much as 25%.

Bottom Line: Organizations should limit the number of hardware vendors in their environments to one or two. When multiple vendors are used, any given system type (e.g., notebook, desktop, tablet) should be kept with a single vendor, creating pockets of homogeneity to reduce the cost impacts.

Business Impact: Enforcement of standards is one of the most important tools available to IT groups in controlling end-user platform costs.

META Group originally published this article on 19 January 2004.