Andrew Bartelsand Brownlee Thomas, Forrester AnalystsWhat advantages can be reaped from using two suppliers for international telecom services? Our policy has been to select one vendor for software applications and other commodity-like products. Are there reasons to follow a different practice for international telecom services?For most technologies that a company acquires, single-sourcing often is the best choice because concentrating purchases with a single vendor will generally provide the best deal, minimize contract management and technology support costs, and reduce complexity. However, there are times (and technologies) when acquiring the technology from two different vendors is the better choice because it provides a fall-back option should one vendor fail to perform and because competition from the two vendors for new incremental business provides ongoing negotiating leverage. The choice for single-source vs. dual-source depends on the risks of a key vendor failing to perform, and the costs and the additional effort\/resources of changing from one supplier to a different one. In turn, these risks and costs largely depend on whether the product or service is commodity-like or more differentiated. Because commodity-like products or services have low barriers to change, and so are generally offered by a handful of large vendors that may be vulnerable to competitive pressures, using dual-source suppliers can provide a good balance between lowest cost and minimizing the risk of vendor problems.Examples of commodity-type products where dual-sourcing is appropriate are voice telecom services and Wintel PCs. Dual-sourcing in this case would provide enhanced flexibility to get the best price, protect against vendor failure or get the best service-level guarantees. There don't appear to be many downsides for commodity-like products or services. This also could be extended to most transport-centric data networking services as well. The key is their substitutability and the relative ease of migrating from one provider or vendor to a different one. When it is relatively easy to change suppliers, dual-sourcing usually offers immediate benefits in terms of reduced cost because one provider can be easily changed for a different one for any new business. By contrast, differentiated products like enterprise application software are almost always best sourced from a single vendor because there needs to be a good match between the software application - in terms of features, functions, platform and integration - and the business process it supports, and the enterprise IT environment in which it operates and to which it needs to be integrated. Having said this, it's important to keep in mind that most companies in financial services, as well as other services industries, generally are not standardizing on a single application suite vendor for all their enterprise applications - mainly because there are no enterprise resource planning (ERP) vendors that are strong enough in core financial services applications or core services transactions systems. With respect to international telecom services, managed services like private IP virtual private networks (VPNs), etc., are not commoditized and are best bought from one vendor, at least within the same geographical region. Most of our multinational financial services clients tend to dual-source commodity-type telecom services on a regional basis. They also often use a primary multiregional service provider like Equant for managed services. These arrangements commonly are supplemented by using new and smaller providers for growth traffic and on high-volume competitive routes to benefit from sharply lower prices or competitive service innovations. Databases and servers fall in between these two ends of the single-source vs. dual-source spectrum. Some of these infrastructure products are more commoditized than others but may not be readily substituted because of data formats and the cost of support\/training for a given infrastructure. Forrester has spoken with some large financial services companies that continue to support multiple database solutions - e.g., Oracle and IBM - and a couple different servers as well - HP and IBM - because they believe they have the ability to shift new buying power\/clout from one vendor to the other in order to get a better price, etc. However, these dual-sourcing arrangements usually are inherited, as a result of either a merger or acquisition or historical fragmented buying practices. As a result, the company already had invested in internal training and support for the different systems. Our experience is that most financial services companies are now trying to single-source their servers and databases - specifically in order to reduce the cost of training employees on different systems and also to eliminate the costs associated with supporting different\/multiple systems. RecommendationsWhile single-sourcing of a technology is the right choice in most cases, for commodity-like IT goods and services dual-sourcing can produce a better balance between lowest cost and mitigation of vendor-related problems. Companies should vary their IT sourcing strategies depending on the nature of the technology being acquired. With respect to telecoms, value-added or managed services appear to be an appropriate service set to single-source. Thus, we generally recommend that a company considering outsourcing its wide area network (WAN) use a single global network operator or, alternatively, an IT facilities-management outsourcer or systems integrator (SI) to manage the provisioning of these services on a CPE-to-CPE (customer premises equipment) basis, including tail-circuits and associated service-level guarantees. Unfortunately, the managed services space is fragmented, so it becomes difficult to achieve both lowest cost and best service by single-sourcing.