I'm being told that if the quote for an FFP purchase is for a large amount of items, like over 100 line items, instead of completing a price analysis on each and every item, we can take a random sample of 20% of the items and only complete a price analysis on those? I don't think this is true but I can't find anything that says it is or isn't. Anybody know anything about this?

See the Contract Pricing Reference Guides, Volume 2, Quantitative Techniques, Chapter 3, which discusses the use of sampling in the conduct of price analysis. It's not as simple as 20 percent.

https://www.dau.edu/sites/default/files/tools/CPRG-Volume-2.pdf

 

GAO denied two bid protests involving sampling. Both appear to be good precedents, and are worth studying to understand what GAO looks at when it evaluates protests of price analysis and sampling, and how these two agencies successfully used sampling. Five rules to remember: 1) Agencies have discretion as to price analysis methods (i.e. GAO unlikely to second-guess the method chosen, unless it tilts the competition or is illogical), 2) Follow the FAR and choose logical methods, 3) Give clear notice in the solicitation how you'll evaluate, 4) Do what you say, 5) Document each step.

Crown Point Systems (2017), https://www.gao.gov/products/b-413940%2Cb-413940.2. GAO denied protest where the agency random sampled 59 of 1441 line items for price realism, because "... the RFP provided that unbalanced pricing would be assessed based on random sampling, not a direct analysis of all prices. ... In this regard, the agency's application of random sampling and the conclusions drawn from that sampling appear both reasonable and consistent with the methodology set forth in the solicitation. The agency provided a detailed description of its methods and results as well as significant underlying documentation, which support the reasonableness of the evaluation. See Business Clearance Memorandum at 17–22 and Price Analysis at 73–167."

Reli Group, Inc. (2019), https://www.gao.gov/products/b-418005. GAO denied protest where "the agency evaluated prices in two separate analyses, performed by two different individuals. First, the agency analyzed the realism of [successful offeror] Fed Pro’s proposed level of effort. To this end, the TEP Chair compared Fed Pro’s level of effort with the level of effort contained in the independent government cost estimate (IGCE), and found the respective levels of effort to be comparable. AR, Tab 14, TEP Chair Business Proposal Evaluation (Fed Pro), at 2. Second, the agency examined Fed Pro’s direct labor rates. Specifically, the FMS compared the direct labor rates for a sampling of Fed Pro’s proposed labor categories with the median labor rates for comparable position titles from the salary.com website. AR, Tab 12, FMS Business Proposal Evaluation (Fed Pro), at 2. Based upon this analysis, the FMS expressed no concerns related to Fed Pro’s proposed direct labor rates, concluding that Fed Pro’s direct labor rates were reasonable and “on par with the median rate.” Id. While not stating that the rates were realistic, the FMS’s analysis found that only one of the proposed direct labor rates for the labor categories sampled was lower than the comparable direct labor rate from salary.com. Id. Specifically, the proposed direct labor rate for Fed Pro’s Project Manager labor category was found to be [DELETED] percent lower than the comparable salary.com rate. Id. In the SSD, the contracting officer found that “[Fed Pro’s] costs [were] realistic based on the requirements of the work and their technical approach.” AR, Tab 17, SSD, at 2. In response to the protest, the contracting officer further explained that both the TEP Chair’s realism analysis and the FMS’s price analysis were relied upon when making the determination that Fed Pro’s proposed price was realistic. ... [and based on this] We find no basis to question the agency’s price realism analysis of Fed Pro’s proposal."

Bottom line: GAO says sampling is permissible so long as there's a clear plan, properly noticed, executed, and documented.