This article is an excerpt from The Art of Tendering: A Global Due Diligence Guide, which is available for purchase.
In its May 2022 decision in Office Depot, LLC, the U.S Government Accountability Office (GAO) ordered the government to redraft and reissue its solicitation after determining that gaps in the government’s historical volume data led to a distorted price evaluation. The dispute dealt with a Request for Quotations (RFQ) issued by the U.S. General Services Administration (GSA) for the establishment of a blanket purchase agreement (BPA) to support GSA retail operations at eight U.S. Army and U.S. Air Force locations.
The complainant alleged that the RFQ’s technical and price evaluation factors were unreasonable since they were based on a market basket of representative items that failed to reasonably account for the government’s actual requirements. The GAO agreed. While the GSA attempted to develop a representative market basket of the 200 most frequently purchased items to evaluate the pricing of competing bidders, the GAO determined that the market basket was not reasonably representative of actual requirements since it was based on incomplete historical data.
While the GAO accepted that the quantity of actual orders may be uncertain when a BPA is established, and that different methods can be used to evaluate pricing based on a representative sample of estimated volumes, it ruled that the method used to evaluate that pricing cannot produce misleading results when assessing the cost of competing bids:
There is no dispute that the actual order quantities of particular items may be uncertain when a BPA is established. Here, the government is establishing a BPA under which tens of thousands of potential items can be ordered. This uncertainty about what items will be ordered can present difficulties for agencies trying to decide which vendor presents the best value when considering price and technical features. This challenge is essentially the same as the one the government faces when awarding an indefinite-delivery, indefinite-quantity contract; the terms are indefinite because the exact times or quantities of future deliveries are unknown at the time of award. FAR 16.501-2.
Agencies have developed a variety of methods or strategies to address this difficulty, including the use of estimates for various quantities of units to be purchased under the contract, Creative Info. Tech., Inc., B-293073.10, Mar. 16, 2005, 2005 CPD ¶ 110 at 3; the use of sample tasks, FC Bus. Sys., Inc., B-278730, Mar. 6, 1998, 98-2 CPD ¶ 9 at 3-5; hypothetical or notional plans that are representative of what requirements are anticipated during contract performance, Aalco Forwarding, Inc., et al., B-277241.15, Mar. 11, 1998, 98-1 CPD ¶ 87 at 11; and hypothetical pricing scenarios reflecting various cost or price eventualities, PWC Logistics Servs., Inc., B-299820, B-299820.3, Aug. 14, 2007, 2007 CPD ¶ 162 at 11-15.
Underlying each of these methods is the central objective of evaluating the relative total cost or price of competing proposals in order to provide the agency’s source selection authority a meaningful understanding of the cost or price implications of making award to one or another concern. MAR, a Div. of Oasis Systems, LLC, B-414810.5, July 26, 2018, 2018 CPD ¶ 266 at 7. It is axiomatic that the agency’s price evaluation method must produce results that are not misleading. Aalco Forwarding, Inc., supra. at 11. Based on the record produced by the agency, we cannot conclude that the government’s sample market basket reasonably reflects the likely estimated breakdown between hardware/industrial and office supply requirements, and therefore, the market basket has the potential to produce a misleading result.
The GAO found that the market basket used in the RFQ was based on a comparatively small sample of the overall requirements that were skewed towards certain categories of hardware and industrial items and that this market basket failed to properly account for other office supplies due to the lack of historical data for those other categories:
Here, the agency has attempted to address the uncertainty associated with future BPA orders by evaluating quotations based on a comparatively small sample of items–200 of the tens of thousands of items that can potentially be ordered under the BPA. As explained above, the agency selected these items because it found them to be representative of the items that are likely to be ordered in the highest quantities. The agency reached its conclusions about what items are ordered in the highest quantities by considering historical market data about item purchases. There is nothing inherently unreasonable about consulting historical sales in order to forecast the government’s likely future requirements.
The record, however, reflects that the 4PL BPA historical data considered by GSA for the purpose of creating its representative market basket was skewed heavily towards hardware/industrial items. When constructing its market basket, GSA’s market research for hardware/industrial supplies included significant sales information for each of the eight locations. Specifically, GSA reviewed historical sales information for the five locations with current hardware/industrial 4PL solutions, and other sales data for the remaining three locations. Thus, GSA had a reasonable basis to forecast its customers’ future anticipated hardware/industrial supply requirements.
In contrast, GSA’s consideration of historical sales information for office supplies was severely limited; specifically, data about the government’s history of purchasing office supplies is nearly absent from the 4PL BPAs used to perform the agency’s market basket analysis.
…Thus, while the agency had a reasonable basis for forecasting estimated quantities for hardware/industrial items at all eight covered locations, it had no (or virtually no) basis for forecasting estimated quantities for office supply items, given the absence of sales data for five locations, and limited sales data from two additional locations. There was only one location, JBER, where the agency had comparable historical data (i.e., 4PL BPA data) available for use in forecasting estimated quantities for office supply items.
The GAO determined that “the agency’s proposed market basket for this procurement does not appear to be reasonable” since the “GSA’s underlying market research relying on past sales was incomplete and failed to reasonably account for the materially different and enhanced office supply offerings contemplated under the RFQ.” The GAO therefore sustained the protest and ordered that the RFQ be amended and re-tendered.
As this case illustrates, scoping a contract to include multiple price categories with unknown future quantities creates significant evaluation challenges. While bid price calculations can be based on a representative sampling of anticipated future needs, the data used to create that representative sampling needs to be reliable to enable an accurate assessment of competing bid costs. When historical volume data serves to anchor a representative sampling, the scope of a contract should be limited to those categories that are supported by that data.
