May the Best Product Win: A Government Procurement Conundrum
Saturday, March 20th, 2004Usually governments try hard to appear fair in procuring ICT (Information and Communication Technology) solutions and services. In response to some industry giants and influential associations, government authorities in the US, Europe and Asia have sought to be fair by sitting on a fence of indecision, avoiding policy pronouncements favoring Open Source vs proprietary software.
Fence sitting has its benefits. In the short run, no one will get fired and the status quo is preserved. But it hardly counts as “fair”. In the long run credibility and transparency will suffer. In developing countries, economic vitality is at risk as well.
But there is progress. We are seeing a movement toward procurement language in government tenders that tries to avoid specifying actual product brands. So far few ICT procurements actually practice this level of functional specification in either hardware or software buys. However, even if procurements adopt a strictly functional approach, is this enough?
The problem is that even cost conscious, functional specifications in ICT procurements remain fundamentally flawed. Two issues plague complex ICT procurements. First, the dynamically evolving nature of technology inhibits adequate specification of requirements. That is functional specifications are hard to write. Second, cost evaluation criteria in procurement specifications are often inadequate. For example, winning a procurement because of low cost can be disadvantageous to the buyer as well as to the economy at large. The economic and social effects of major ICT solutions work well beyond the perimeters of individual ICT procurements. For both the purchaser and society, low cost may achieve the least benefit.
Functional Specs are Hard to Write
Support of interoperability and data standards is an essential basis for fair procurements. Unfortunately, requirements for legacy system compatibility or minimizing re-training costs or mandating selective but relatively obscure features, almost guarantee the elimination of competing proposals from the Open Source community. Cost containment and familiarity requirements such as these are on the surface quite reasonable but tend to defeat impartial functional specifications.
A more intrinsic problem with standards-oriented, functional specs involves imposing standards prematurely. Successful standards usually evolve over many years and for non-trivial systems, continue to evolve. Furthermore, standards and their complex interactions are never known fully in advance. That is, software procurements typically cannot be specified in sufficient granularity or with adequate foresight to anticipate the working requirements of most large-scale projects.
Furthermore, many important features may not be interoperable. For example, data standards may degenerate to least common denominator compliance where embrace and extend strategies can seriously twist the protection standardization might otherwise provide. Consider recent Microsoft filings for XML-related patents that have the potential to torpedo any genuine interoperability between office document processors. In addition, unique product features not adequately represented in relevant standards may be needed by the purchasing organization.
To help ease the difficulty of developing of functional specifications, several steps may be taken. First, interoperability and adherence to current standards must be demonstrated. Then, for additional features and technologies, which address evolving standards or vendor “improvements”, the successful vendor must be required to submit its proposed solution and its component technologies to relevant standards bodies as appropriate. This will help defeat embrace and extend forms of vendor lock-in.
Low Cost Achieves Least Benefit
Complementary to using adequate functional language in specifications, large-scale procurements should consider “best value” instead of “low cost” evaluation criteria. Low bidder victories often fail because of dishonestly low bids and subsequent cost overruns. Also, low cost criteria tend to inhibit innovative bids. When large-scale government procurements are evaluated against “best value” criteria, the low-cost bid is not automatically chosen but, instead, a range of desired benefits are measured. In this way, choosing an innovative solution might save money through new system efficiencies in contrast to selecting a low-cost bid that maintains an antiquated or low efficiency business process. For example, an open source solution might offer low maintenance diskless workstations whose software is strictly tailored to the job a user is supposed to be doing. This can result in using smaller, cheaper machines that do not have the ability to be misused by being loaded with typical work-place distractions, such as games and music. Therefore, the potential cost of retraining faced by an open source solution might be absorbed by greater workplace efficiency.
In addition to using best value evaluation criteria for complex ICT procurements, the procurement process must be protected by a policy that strives not only to be vendor and technology neutral, but addresses larger issues of fair trade and social benefit. By choosing appropriate policies, a larger scope of benefits can be supported across various ICT procurements. Additional weight can be given to bids that address issues such as promotion of vendor competition, increase of domestic skills and jobs, and control of security. Evaluation criteria must consider the larger benefits of open source such as restraining the disadvantages which proprietary technologies can impose on a national economy over the long run. Unfortunately such larger issues are usually beyond the scope of individual procurements and must be introduced at a government policy level.
“May the best product win” can be a screen for simply preserving the status quo. Today the status quo is all about vendor lock-in, escalating licensing costs, loss of domestic skills and jobs, increased pressure to constrain piracy, and loss of control of security. By implementing fair ICT procurement practices informed by a government policy that promotes the larger economic welfare and social benefits, we can begin to rephrase “May the best product win” into “May the most beneficial product win”.
