Improving the Procurement Process:
National Computerized Election Systems
by Jay Cornish

Overview
Computer information technology integration projects can be complex, time-consuming, risky, resource-intensive - and critical to your election management organization's future. How do you ensure their success? Start with proper procurement methods. This paper offers a supplier's perspective on how national election authorities can improve their procedures for procuring large-scale computer information systems. Specific goals for this process are optimizing the procurement cycle, ensuring a level "playing field," reducing the risk of bidder complaints and protests, and increasing the likelihood that the selected solution will be delivered on schedule and on budget.

The Challenges to Election Authorities
In recent years, numerous national election offices around the world have commenced ambitious information technology (IT) modernization projects. Whether it is a new voter registration system, voter identification system, voting system, tabulation system, and/or results-reporting system, these projects tend to be large-scale, complex procurements. To be successful, careful planning is needed to orchestrate not only the competitive bid for the new technology, but also the project's financing, the legislative impact of the new technology, the public education campaign, the political consensus-building, and so forth. Compounding these challenges is the zero tolerance for slippage in the implementation schedule. While other government agencies enjoy the relative luxury of setting their own IT procurement and implementation deadlines, an election authority is forced to work with the deadliest of deadlines - namely, Election Day. Given these significant challenges, it is not surprising that these modernization projects far too often produce unsatisfactory results - or even worse, fail. Exacerbating the challenge even further is the fact that while there are excellent books and articles publicly available to vendors on "how to bid," most of the material on "how to procure" is only available in the form of governments' rules and regulations. Especially in developing countries, where the procurement of large-scale IT projects is a relatively new process, there are few if any proven IT procurement models and methodologies. [TOP]

A Systems Integrator's View
Unisys believes that, regardless of the application or the country, the best managed procurements share key common characteristics. To reduce risk, large-scale IT projects are often contracted to "systems integrators," who in turn subcontract specialized technology companies to supply the voting equipment itself and expertise in elections administration. Systems integrators are large companies experienced in managing multi-vendor projects that require seamless integration of hardware, software, networking, services, and long-term support. The primary benefit of contracting a systems integrator is that the user works with one vendor whose responsibility is to manage the subcontractors and deliver the overall solution on schedule and on budget. A qualified systems integrator has the technical expertise, the proven delivery methodologies, and the project management skills necessary to get existing disparate systems talking to one another, to get new equipment working in tandem with old, and, in parallel, to keep the organization operational in the process. [TOP]

Six Steps to Successful Management of Systems Integration Projects

2. Choose Carefully When evaluating suppliers, seek those with proven experience and depth of technical resources. If using an integrator, ensure that all of the integrator's subcontracted partners are also highly experienced "best-of-breed" technology suppliers. Does the integrator have a large enough local presence to implement and support voting sites throughout your country? Is your staff compatible with this long-term partner?

3. Start At Top It is critical that the highest-level officials within the election office understand the project and are committed to its success.

4. Talk Early and Often Ensure that staff are working in tandem and sharing knowledge.

5. Beware of Scope Creep Integration initiatives tend to grow. At the outset, collaborate with your supplier to identify who is responsible for what tasks, and establish measurable parameters to keep the project on time and on budget.

6. Frequently Evaluate Progress Set deliverables milestones so that you can measure the project's continual progress, evaluate the supplier's work, and ensure that all of the primary goals are being met. [TOP]

IT Requires a Different Procurement Approach
Traditional government procurement models typically do not work well with IT. For example, the initial purchase price of IT is often far less important than other factors such as total cost of ownership and the relative fit within the strategic direction of the entire IT environment. IT systems are advancing at such a rapid pace that often by the time a government has procured and implemented a system, it has become obsolete. Furthermore, most large-scale IT projects are highly customized, where the vendor must design a solution that meets a unique set of requirements. In order to prevent "requirements creep" (i.e. uncontrolled requirement changes that adversely affect the budget and schedule), the customer and the contractor must collaborate closely to define change control mechanisms. But regulations often act to minimize the communication between buyer and seller during the procurement process. While governments have made progress in streamlining procurement procedures for IT commodity purchases (such as personal computers), similar procurement procedures are often ineffective when applied to systems integration procurements. For best results, look for two key indicators when selecting a supplier: a) a history of success delivering similarly-scaled projects on time and on budget, and b) a proven set of systems integration methodologies and tools.

Further, a major IT project's success or failure is often determined early in the process, when you identify a supplier. The supplier's team must be compatible with your IT staff on both a personal and professional basis. The early stages of implementation often require intensive collaborative requirements and design planning sessions that can last for many weeks. Again, it is critical that the supplier possess a solid background in election systems management in order to thoroughly and unambiguously understand your objectives. Major IT projects fail more often due to human problems - politics, personalities, unclear specifications, unmet expectations, unmanaged risk, poor communication and planning - than due to the technology or budget. [TOP]

The Outsourcing Option
One of the delivery options typically offered by IT firms is "outsourcing." Outsourcing occurs when a national election authority contracts with a systems integrator to assume responsibility for one or more IT functions, characterized by the transfer of assets that can include the data center, wide-area networks, software applications development and maintenance functions, and end-user computing. Outsourcing IT systems may be of particular interest to national election authorities in developing countries that have experienced technical "brain drain" - i.e., well trained IT professionals leaving the public sector for more lucrative private-sector jobs. Outsourcing non-core functions is the embodiment of a basic concept: Do what you do best, and engage experts to do the rest.

An election authority that is considering the outsourcing option must articulate the value of doing so. Is it to cut costs? Is it to obtain technology skills? Incorporating outsourcing into your overall strategy allows your organization to concentrate on continuity and growth. Outsourcing enables your organization to re-deploy critical resources to activities that are more in line with your organization's core competencies. To be effective, any outsourcing RFP should ensure that personnel considerations are addressed. Service level agreements, key deliverables, and schedules must be clearly delineated and managed. Equally important are contractual terms such as the performance measurement and termination clause, which will allow you to manage what was proposed, and to take the necessary action in the event of non-performance by the supplier. [TOP]

Start Early with Functional Requirements Analysis
The procurement cycle for a typical large-scale technology project can take a year or longer, from concept to contract award. From contract signing to final system acceptance, implementation can also typically take at least a year. Therefore, the formal procurement process should begin at least two years prior to the Election Day when the system is to be first used, and it is recommended that a contingency period be built into the schedule.

A logical first step is to differentiate between the election system processes that you are considering modernizing. There are clear distinctions between voter registration, vote casting, vote tabulation, and vote reporting systems. You may wish to procure these projects separately, or combine a few together, such as the casting and tabulation systems. However, it is recommended that the registration system be modernized first, and procured separately from the other election systems. Once the voter database has been updated and "sanitized" for duplicate and deceased records, the other election processes will flow more efficiently.

Every national election authority appears to have a different set of compelling reasons to invest in computerization. For some nations, the fundamental problem can be voter fraud at the polls or corruption during the manual counting of paper ballots. For other nations, the fundamental problem can be the need to increase the percentage of registered voter participation; reduce the cost of elections; increase the efficiency of elections management; or improve convenience for voters. Whatever the primary problem, it is important that the functional requirements for the new system focus directly on solving it. During the needs analysis discussions, make sure that every requirement is tied to solving the fundamental problem and that it is as unambiguous as possible. This clarification process serves as an important preventive measure for overall project cost containment because the cost to fix false assumptions in the requirements definition phase increases almost exponentially during later phases: [TOP]

Assign Procurement Committee and Consultants
Once the project has been identified, the next step is to assemble a multidisciplinary committee to refine the functional requirements analysis, develop the Request for Proposals (RFP), evaluate bids, recommend a selection, negotiate the contract (this is generally handled by someone in the contracting/legal group), and co-manage the implementation. This team should consist of people familiar with the functional organization and the overall long-term direction of the organization. It should include a full-time dedicated Procurement Manager and representatives from the end-users group, the technology group, and the contracting/legal group. This team should be kept intact throughout the entire procurement process. If people with specialized expertise are unable to participate on the team, external consultants may be an appropriate substitution. Seek consultants that have proven and successful experience assisting other federal government agencies procure similar systems and that are not biased toward a particular supplier or solution. [TOP]

Define the Problem, not the Solution
The functional requirements analysis should be focused 100 percent on defining the problem, as opposed to exploring solutions to it. The bidders - especially systems integrators - will customize the best possible solution for your particular set of needs. Start with a detailed written description of existing flow and management of information within your organization. Then prioritize the existing problems with that process and your objectives for improvement. A typical statement of requirements will include a description of the existing system, quantification of volumes of throughput and sizes of files, remote sites, and a statement of expected benefits. [TOP]

If You Have Ample In-house IT Resources...
Some government agencies have reduced the risk of failure of large projects by applying the general engineering principle of a "Black Box." Each Black Box is a discrete component, such as the database or the network, that is both independent and integral to the entire system. A Black Box can be a commercial off-the-shelf product or service. The customer acts as the systems integrator, creating a common, open, scalable architecture, and then procures the Black Boxes from various specialized vendors. This common architecture must cover everything from operating systems to networks to database management. If your election organization has a large, stable in-house IT team that is both familiar with a wide range of available technology and capable of managing the integration of a large computer systems project, then the Black Box approach may make sense; otherwise, you will likely need to select a systems integrator partner. Systems integrators utilize implementation methodologies based on similar approaches to the Black Box approach, subcontracting best-of-breed specialized technology suppliers. [TOP]

Phased Implementation
In order to reduce risk and manage change, it is recommended that any national election authority that is planning to put in place a new computerized system of any type do so with a phased implementation approach. For example, if the objective is to replace paper ballots with "electronic ballot boxes" or "direct recording equipment," then the first phase of national implementation could be to do so only in cities over a certain population size. Alternatively, the first phase could be implemented for all voters living in a particular region of the country. The second phase of implementation (not necessary a completely new procurement) could then be for cities with fewer voters, or for additional regions. This phased implementation approach could be applied to the introduction of any other new election technology, such as replacing manual counting of paper ballots with optical scanning/tabulation equipment, or utilizing the Internet to interactively disseminate election results. [TOP]

Specify Open Systems
Further, it is recommended that you not dictate specific technical platforms, architectures, hardware, or protocols unless your organization has some sort of compelling long-term commitment to certain standards, and your organization will be responsible for maintenance of the delivered system. The trend today is away from proprietary systems and toward "open systems" that comply with international communications and interface standards, and are modularly scalable - i.e., cost-effective add-ons do not require replacements to the original investment. Having the option of choosing among a number of suppliers is usually the most cost-effective situation when undertaking procurements; this is especially important with computer systems, since they can become obsolete without continual attention to upgrading. It is also important to mandate that this "mission critical" system have an extremely high level of availability, or uptime, during the Election Day process. Your contingency requirements may include total redundancy of core data and key components in order to minimize the risk of downtime during critical periods of use. [TOP]

Cost-Benefit Analysis
To justify funding and/or financing, the procurement committee will likely be required to compare the costs versus the benefits of a modernization project. This analysis should include benefits such as:

The analysis of costs should not only include the payments to the vendor but also involve the costs of a national voter education campaign to prepare voters on new procedures, as well as the internal costs associated with re-organization, training, and change management. Further, the cost/benefit analysis should assess the technical, financial, contractual, resources, and system obsolescence risks that may be involved in the project. [TOP]

Appropriate Funding
Once the steps outlined above are complete, you will likely then have enough functional requirements information for internal and external funding and/or financing organizations to evaluate, justify, and appropriate funds and/or financing for the project. Financing organizations such as the export-import banks and the development banks will require your organization to comply with their rigorous procurement procedures in order to qualify for their assistance. Be prepared for these important procedures to add months to the cycle. [TOP]

Issue the RFI
Prior to issuing a Request for Proposal, it is recommended that you issue a Request for Information (RFI) document to any and all vendors that may be interested in bidding on the project. The RFI alerts the vendors that your organization is in the information gathering stage but is not yet ready to make a purchase. This is an opportunity for your organization to receive a tremendous amount of valuable consulting information; determine if solutions exist at reasonable costs; and, qualify the bidders for the upcoming RFP phase. The RFI should be a summary of your functional requirements document, with a clear description of the present system, your objectives, and your expectations. You can ask for price estimates, but be prepared for very conservative quotes from vendors at this stage. Invite vendors to visit your organization to obtain a hands-on understanding of how it is presently functioning. An RFI also helps eliminate surprises for the vendors in the upcoming RFP phase of the procurement cycle. The RFI should also request that vendors supply qualifying information about their company's:

You will need the same sort of information about any subcontractors that the prime contractor plans to involve in the project. Based on the responses to the RFI, you will be able to narrow down the list of qualified vendors that you will allow to respond to the RFP, thereby making your evaluation of the RFP more efficient and meaningful. [TOP]

Issue the RFP
A Request for Proposal represents a formalized process for documenting a procurement; allows for evaluating different solutions; and provides a means for monitoring the performance of the winning supplier. Issuing a poor quality and ambiguous RFP can seriously damage an election authority's strength and effectiveness. A poor RFP can lead to selection of a solution that is significantly higher in cost, take longer to implement, and unlikely to meet your goals. The typical RFP for a large scale IT project includes a Technical Section, a Management Section, and a Contracts and Pricing Section. These sections are interconnected and mutually dependent on each other. Your government contracts and procurement office likely has a set of guidelines for the legal "terms and conditions" and formatting of the RFP. You can stipulate in the RFP that the supplier's proposal will become part of any subsequent contract awarded, thereby setting the foundation for contract negotiation. It would take many pages to describe the details of a model RFP, so suffice it to say that, at the very least, an RFP should be a thorough, unambiguous, and quantitative description of your functional requirements and goals. Typically an IT RFP is issued to five to ten pre-qualified bidders, and the deadline for proposal submission is typically eight to twelve weeks.[TOP]

Basic Sections of an IT RFP

Benchmark Testing
Benchmark testing hardware and software systems is a means for your procurement committee to view prototypes and real-world operations of the technical solution that you are procuring. There are many existing standards and methods for testing. For example, in the United States, the National Institute of Standards and Technology (NIST) publishes test standards, and the U.S. Federal Election Commission has published standards and requires the use of independent laboratories to test the performance of voting equipment. Whenever feasible and appropriate, it is recommended that tests be conducted on the supplier's existing installed customer systems. [TOP]

Evaluation
The purpose of the evaluation of the proposal and benchmark test is to assess the supplier's capability to successfully perform the contract. Your evaluation will need to follow strict scoring guidelines so that the same objective, measurable criteria are used when evaluating each supplier. Your guidelines are essentially a weighting system, assigning value to criteria such as the supplier's understanding of the functional requirements, compliance with mandatory requirements, qualifications and references, oral presentations, project management plan and implementation schedule, benchmark test results, and pricing. (A word about pricing: It is well documented that government agencies that have already implemented multiple IT projects tend to weigh price less heavily as an evaluation criteria in their evaluation than do agencies that are procuring a system for the first time.) [TOP]

Contract Award and Negotiations
Once you have selected the winner, allow for a few months of contract negotiations. Also, no matter how much you do to prevent and discourage protests, be prepared for disputes from the losers. You can diffuse complaints by being accessible and openly explaining and clarifying any and all concerns. Suppliers do not want to have to protest an award, so give them every reason not to. [TOP]

Win-Win-Win
By applying these "best practices" to your next procurement process, your election office will likely achieve win-win-win results:

Selected Readings

Dr. Robert Ambrosino. Managing Projects in the Information Age. Sacramento, CA: Government Technology Press, July 1998.

Michael Asner. The Request for Proposal Handbook. Sacramento, CA: Government Technology Press, 1995.

Michael Asner and Michael Moss. Handling Supplier Complaints and Protests. Sacramento, CA: Government Technology Press, 1998.

Kimball W. Brace. Developing a Statewide Voter Registration Database: Procedures, Alternatives, and General Models. Washington, DC: Federal Election Commission, 1997.

Paulo César Bhering Camarão. Electronic Ballot: Democratic Legitimacy. Sao Paulo, Brazil: Empresa das Artes, 1997.

Computerizing Election Administration, Volumes I, II, and III Washington, DC: Federal Election Commission, 1987.

Donald C. Gause & Gerald M. Weinberg. Exploring Requirements: Quality Before Design. New York: Dorset House Publishing, 1989.

William J. Hirsch. The Contracts Management Deskbook. New York: American Management Association, 1983.

In Search of Integration Gratification: Does Your Integrator Measure Up? Framingham, MA: Computerworld Magazine, July 28, 1997.

Harry Neufeld. "Computerizing Electoral Administration, Part I and Part II." Washington, DC: Elections Today (IFES), May 1995 and October 1995.

Tod Newcombe. Outside The Box: Engineering Concept Aids Agency. Sacramento, CA: Government Technology Magazine, July 1998.

Bud Porter-Roth. How To Write A Request For Proposal. Silver Spring, MD: Association for Information and Image Management, 1991.

Roy G. Saltman. Adopting Computerized Voting in Developing Countries: Comparisons with the US Experience. Palo Alto, CA: Computer Professionals for Social Responsibility, 1998. [TOP]