The Department of Homeland Security recently canceled a large-scale project attempting to make agile software development services easier for agencies to buy, and the leader of the Digital Service team that worked on it is searching for lessons in its failure.
DHS launched Flexible Agile Support for the Homeland, or FLASH, last year. The contracting vehicle pre-approved 13 vendors that passed a technical development challenge to sell their services to DHS agencies. Shortly after its launch, progress was halted by protests multiple times from multiple companies that didn’t make the cut and felt the selection process was unfair. Last month, DHS moved to cancel the contract, still dogged by protests, citing concerns that it had not had enough information to “reasonably evaluate the offerors.”
DHS’ three-year, $1.5 billion contract was partially modeled after a similar effort at the General Services Administration. GSA’s tech consultancy 18F launched the Agile Blanket Purchase Agreement in 2015, which also pre-approved agile vendors, and DHS officials had adopted some of their practices when crafting FLASH, including evaluating them based on a technical development challenge instead of a lengthy proposal.
» Get the best federal technology news and ideas delivered right to your inbox. Sign up here.
Eric Hysen, the Google transplant who was formerly executive director of DHS’ Digital Service team, took to Medium this week to reflect on FLASH's failure. Its cancellation was the “biggest disappointment of my time in government," he wrote, with "the entire effort being canceled over a year in, thousands of hours of work from dozens of people reduced to a few pages of legalese on an obscure government agency’s website.”
Among DHS' mistakes: launching FLASH too quickly, raising the stakes with a $1.5 billion award amount, and attempting to evaluate too many offerors, Hysen wrote.
His team had tried to create a system that let agencies depart from the old purchasing process of months and occasionally years of solicitations, proposals and bids. A prolonged acquisition process requiring hundred-page proposals might work for helicopters, but an agile, piecemeal buying method is better for software, Hysen wrote. Instead, companies interested in FLASH came to DHS and built a simple product on site. DHS also recruited companies that had never worked for government before.
But in practice, the new system proved to be too complex to succeed.
"To start, we went way too big with FLASH, way too quickly," Hysen wrote. Though the $1.5 billion award amount attracted more than 100 bids, "we weren’t ready for it."
And the teams at DHS didn't have experience evaluating the technical challenges, so "they didn’t scale the way we thought they would. We should have put a much lower dollar value on FLASH to start and run a larger contract later on."
DHS should also "never have gotten into a position where 100 companies were participating in half-day technical challenges," Hysen wrote. They were a "logistical nightmare" that required opening a second location and ushering more than 1,000 people in and out of buildings for three weeks.
Sometimes his team scheduled contractors' challenges with little notice, "which led to some feeling they had inconsistent experiences. Multiple companies were participating in the challenge at once, meaning our teams were literally running between different conference rooms rather than focusing all of their attention on one team."
While the technical challenge let DHS see the companies practice agile methodology and evaluate them based on their performance, not their proposal, 100 teams were too many to manage, leading to eventual protests on FLASH. After the first round of protests, DHS re-evaluated bids and whittled down the list to 11 companies. The agency immediately faced with a barrage of new protests.
Hysen wrote he expected the protest process to continue but to eventually be resolved, which "made it all the more shocking" when DHS canceled FLASH after he'd left his government role. Still, "I ultimately think it was the right call," he wrote. "We screwed up, and the mistakes were serious enough that correcting them would have been impossible or turned FLASH into something that no longer did what it was supposed to do."
Though DHS could have resolved the protests by simply awarding those companies a spot on the contract, "it would have killed FLASH" because its eventual goal had been for DHS to "run quick competitions against the smaller number of companies ... without needing to do their own deep technical reviews."