Page 32 - FCW, September 30, 2016

WTInsider
4 ways RFP writing
An analysis of five solicitations for contracts worth a total of $7 billion revealed a lack of clarity in agency RFPs
falls short
BY NICK WAKEMAN
It is no secret that solicitations can be hard to read. The language is technical and often convoluted. It can lead to con- fusion and misunderstandings — not to mentions delays and inefficiencies.
But just how bad is it? Well, maybe even worse than many of us thought, according to a report released this summer.
VisibleThread, whose software prod- ucts use algorithms to analyze the lan- guage of documents and websites for updates and clarity, turned one of its tools loose on five solicitations for con- tracts worth a total of $7 billion.
The company’s primary business is helping contractors track changes to solicitations. Its algorithms don’t just identify changes in wording but also highlight changes in context and sub- stance. A second product analyzes the clarity of the language used on websites.
It was the website tool that Visi- bleThread used to analyze the requests for proposals for the General Services Administration’s Human Capital and Training Solutions; the Department of Health and Human Services’ Unified Program Integrity Contractor; HHS’ Research, Measurement, Assessment, Design and Analysis; the Air Force’s
Joint Range Technical Services II; and the Navy’s Fielded Training Systems Support IV.
VisibleThread CEO Fergal McGov- ern said the RFPs were evaluated for readability, passive language, long sen- tences and word complexity density. The scores were then compared with results of Flesch-Kincaid readability tests, which the Navy developed in the 1970s to improve training manuals and other technical documents.
Each of VisibleThread’s four catego- ries had a target goal:
• Readability — a score of 50, which is about an 8th-grade reading level.
• Passive language — 4 percent or fewer of the sentences have passive construction.
• Long sentences — 5 percent or fewer have 25 or more words.
• Word complexity density — a score of 100. This scan looks for complex words and phrases based on the plain language guidelines the federal government has established.
VisibleThread did not review entire solicitations but instead focused its anal- ysis on the statement of work, Section L (instructions) and Section M (evalu- ation criteria).
Those areas often cause the most confusion for contractors and result in plenty of back and forth between agen- cies and bidders, McGovern said. After a contract is awarded, those sections are often cited in bid protests and lead to delivery issues.
The company’s analysis revealed plenty of room for improvement:
• The readability score was 32.9, four grade levels higher than recommended for clear writing.
• Passive voice was present in 14 per- cent of the sentences, more than three times higher than recommended.
• Twenty percent of sentences exceed- ed the recommended length.
• The average complexity score was 3.67, suggesting opportunities to sim- plify word choices, according to the company’s report.
Of the three parts of the RFPs that VisibleThread evaluated, Section M (instructions) scored the worst. The statements of work were poor perform- ers as well.
One of the key takeaways is that the quality of solicitations can vary widely from one section to the next, which McGovern said suggests that different people write different sections and no
WashingtonTechnology, a sister publication to FCW, covers all the ins and outs of the IT contracting community. Learn more at WashingtonTechnology.com.
30 September 30, 2016 FCW.COM