Preamble: I'm Kevin Merritt, CEO of Socrata, a leading provider of open data services, based in Seattle, WA. While Socrata is focused squarely on helping government organizations share their data in human-comprehensible and machine-readable ways, I don't personally have a long tenure of experience selling software or services to the government. In full, transparent candor, the procurement process is the most challenging part of working with the government. It's complicated. So much so that many companies selling to the government have GSA contracting specialists to help navigate the process. As a taxpayer and businessperson, I'm in favor of making it simpler and more effective for innovative companies to do business with the government.

In March the General Services Administration (GSA) conducted a pilot project designed to make procurement a more transparent process. In the pilot, GSA submitted to industry an RFI for the next phase of evolution of data.gov. I participated in the pilot and want to share my feedback as a means to hopefully improve the overall procurement process.

Historically, GSA would issue a Request for Information (RFI) and industry would submit their responses in private. There is no back-and-forth communication between GSA and industry while the RFI response window is open. Without regard to GSA specifically, an RFI is often used to conduct basic market research and to help refine the requirements that will ultimately be included in a formal solicitation - the Request for Proposal (RFP).

The BetterBuy pilot project was novel in that it asked industry and interested individuals to submit their responses to the RFI, through a publicly editable wiki. It was the most transparent and collaborative RFI conducted by the federal government to date.

For those unfamiliar with wikis, they are web-accessible, collaborative content management systems. They allow anyone to create new content pages and even to edit the content created by others. Wikis work well for a few reasons:

• General trust between and among the participants.
• A community moderation system that identifies and corrects erroneous content.
• A robust audit trail that shows who created (or changed) what.
• Tools allowing site administrators to manage content revisions and versions.

In general, the BetterBuy wiki pilot was a success. Quite a few people participated and the discourse was civil. I believe the GSA data.gov team received valuable feedback, which will shape and influence the detailed requirements of the RFP. As importantly, I believe industry received some valuable feedback from the GSA data.gov team.

GSA would not have conducted the pilot unless there exists a serious intent to improve the procurement process. I believe their intention is sincere and genuine. To that end, I have some constructive feedback and suggestions for improving the process in the future.

1. Don't map the structure of the RFI to the structure of the anticipated RFP. By doing so, the participant pool is somewhat restricted to those comfortable and/or familiar working within a prescribed format. What might be better is for the soliciting organization to loosely enumerate their questions and requested feedback, and broadly categorize those by subject area - background, physical architecture, logical architecture, security, functionality, accessibility, pricing, terms and conditions, etc.
2. One of the potential benefits in a more open process is a level playing field. As such, provide an RFI platform that offers more transparency about the participants themselves. For example:
1. Participants should use their full names and identify their affiliations;
2. GSA administrators should be easy to identify;
3. Industry submissions, including the possibility for multiple submitters per company, should be easy to identify;
4. I suggest having three contributor profiles with corresponding, visually distinct badges - administrators, industry and interested citizens.
3. The wiki site itself was a little sluggish throughout the response period, but especially in the last few hours leading up to the close of the response window. Make sure there are adequate resources so the site operates smoothly In future acquisition collaboration programs.
4. The GSA itself should be more collaborative during the RFI response window. For example, if industry responds to an initial question in a way that doesn't provide the answer/clarification it's seeking, why not annotate the response with a follow-up question? It's this kind of back-and-forth dialog that represents a really open and transparent discussion.
5. Consider using IdeaScale or similar crowd sourcing platforms instead of a wiki. Potential benefits include:
1. A more natural environment for questions and answers, especially when it's anticipated that there will be multiple answers per question and/or multiple answerers per question;
2. The ability to more elegantly categorize questions by subject area;
3. Greater ease of use, improving participation;
4. A clearer question and answer threading system allowing participants to both respond to official GSA questions/requests but to also reply, rebut and/or provide feedback on the answers provided by other participants;
5. Visual indicators and/or badges identifying participant roles and their affiliations;
6. The ability for participants to vote up/down good/bad ideas submitted by others;
7. The ability for participants to suggest new questions that weren't initially thought of;
8. The ability for administrators or participants to mark a response as the best answer;
9. More transparency around the number of answers provided, new questions asked, top votes, etc. via statistics and analytics.
6. Without regard for my 5th suggestion, if a wiki is the preferred approach for collecting industry feedback, take advantage of wiki "sections" which can be edited in isolation, more locally, without locking the entire page and without encountering edit collisions.