In this installment we consider the current state of government e-procurement systems, paying particular attention to those operating at the state level.

Nearly all of the 50 states now have systems that announce solicitations to vendors. Yet only 10 have systems that process electronic orders. These are Maryland, Massachusetts, California, Michigan, North Carolina, Washington, Virginia, Texas, Wisconsin and New York.

Forty-two states will have implemented or begun implementation of full-blown e-procurement systems by the end of 2003, according to Gartner Dataquest.

States can now expect initial investments of between $5 and $15 million. These costs can go up considerably when adding the costs of tying procurement processes to back-end financial systems, along with the costs of training and customer support.

False Assumptions

In the rush to build the early systems, a number of assumptions were made that, in some cases, turned out to be false. Among these:

• The self-funding model is win/win.
• Agencies will fully support the new systems.
• Vendors will be join in high numbers.

States that are just building their systems now are in a position to learn from these false assumptions. Let's take a closer look.

Lessons Learned

The general consensus now among procurement officials and e-procurement implementers is that the "self-funding" model (where implementing companies generate revenue mainly through vendor fees) is dead. (There are still remnants of it in cases where governments guarantee agency participation or minimum revenue volume.)

A couple of reasons why this model is dead:

• Too many vendors balked when asked to pay both subscription fees and transaction fees. (One system, Virginia's eVA, is dropping vendor transaction fees for 2003.)
  
  
• Transaction volume has been too low to generate sufficient revenue for e-procurement implementers.

Beyond the death of the self-funding model, here are some other lessons learned:

• You can't put a round peg into a square hole. Too many B2G implementers pushed offerings that were not sufficiently tailored to the unique needs of the public sector. For example, government agencies must ensure that e-procurement procedures comply with security rules involving such matters as receipt validation, bid opening and negotiations. E-procurement tools did not always address such requirements.
  
  
• It's not easy tying front-end to back-end. System builders underestimated the importance and difficulty of tying the ordering side to back-end financial systems. The ultimate goal is to integrate to the point that, when a purchase order is created in the procurement office, it is automatically checked against funds availability in accounting. Matching the procurement office's definition of transaction type to the needs of accounting has been difficult.
  
  
• Vendors' needs are diverse. In the private sector, companies can limit supplier numbers to make their lives simpler. That is usually not possible in the public sector. Generally, government systems must contain all the data elements necessary to handle both a company like IBM and the typical small town office supply store, for example.
  
  
• Not enough commitment. Implementing firms complain that there is a lack of commitment on the part of agency decision-makers to build full-feature systems. Part of this stems from the early popularity of the self-funding model. It's harder now for implementers to ask agencies to spend millions up front.
  
  
• Build and they might not come. Vendors need to see some future benefit to invest the time in joining e-procurement systems. The most successful systems seem to be those built on an "agencies-MUST-take-part" foundation, rather than an "agencies-MAY-take-part" foundation. When their buyer customers use systems from the outset, vendors soon follow.

Oregon: Late Adopter

States that have not yet adopted e-procurement systems have the luxury of considering what has worked and what has not. Oregon is one example. State officials have taken a hard look at those that have come before, and they've come up with the following best practice suggestions:

• Know the costs. The ability to aggregate cost data by commodity, service and vendor is an essential first step to forecasting future expenditures, consolidating purchases and attaining reduced costs through volume purchases.
  
  
• Redefine processes to eliminate low-value-add components. Unecessary requisition approvals are easy targets for removal.
  
  
• Establish metrics and measure against them.
  
  
• Delegate low-value, low-risk procurements to agencies. Central procurement resources should be allocated to analyze complex, high-risk, high dollar value procurements.
  
  
• Limit project scope. Isolate the addressable areas of spending and identify areas with the highest product and administrative costs.
  
  
• Seek best value, not just low cost. Governments should change laws that require low-cost-only awards to allow for best-value procurement.
  
  
• Market benefits to top management and key stakeholders to obtain political backing for changes. Involve the agency's controller's office at an early stage because of the need to integrate with accounting systems.
  
  
• Improve communication between government and industry. Release draft bids for public comments. Hold regular briefings. Ensure that contracts clearly communicate expectations and commitments of both parties.
  
  
• Adopt new procurement models. E.g., allow for multijurisdictional contracts such as Western States Contracting Alliance, http://www.aboutwsca.org; establish contract and procurement processes from other governments and proven systems.