The Obama Administration’s plan to fix its glitch-prone HealthCare.gov website by hiring more consultants and bringing in droves of the best and the brightest, risks making the solution team too big to be effective, according to tech workers in the Mission District.
More people means more opinions and more mistakes, or, as tech entrepreneur Doug Simpson, 41, put it “[with a big team,] meeting in the middle just means you make a middling product. Group consensus never builds great products”
It’s a rule most programmers learn early on. Coined by the computer scientist Fred Brooks in the 1970s, Brook’s Law says that while bringing on more people to help meet a deadline seems logical, it can often slow down the process.
New people need to be trained and familiarized with a complex project, which takes time and resources. Additionally, the more people and teams that are involved, the more communication slows, and the longer it takes to achieve consensus on design decisions. In short, too many cooks can spoil the broth.
“With a small team you can be so nimble and really respond to things” said Cristina Mercando, a programmer who helped design the machine-learning and recommendation engine website Hunch.com, which eBay acquired at the end of 2011. “Once we got acquired by a big company it got much harder to coordinate, and you could just feel how things slowed down,”
How would she fix HealthCare.gov? Get one top programmer and put them in charge of a small team of experienced pros to tackle it from the bottom up.
This was a sentiment echoed by many tech experts. At present, there are more than 50 teams working on the healthcare exchange. So many teams disincentivize anyone from taking risks, and make it too easy to blame others for work not getting done, programmers here said.
“You need one manager and one central team. When things get too big, it’s too hard to keep everyone accountable for their work,” said, Ritchie Zeng 20, cofounder of a wearable electronics firm.
His co-founder Nelson Zhang, 19, agreed. “In these sort of contracts, especially working for people who are in government and probably not experts, the incentive to overcharge and under deliver is just huge.”
All of the programmers understood the problems of a start-up going wrong.
“On time, on budget, or with the promised features,” said local technology entrepreneur and startup specialist Brady Forrest. “Pick two. That’s the rule of a product launch. You’re never going to hit all your goals.”
The health care website is only designed to accommodate 60,000 people at a time, yet during its first three days online, it had 8.6 million visitors. Health and Human Services reports that by Saturday the 19th, 19 million people had visited the site.
Though traffic on the site has slowed since its launch and improvements are being made, the website still stalled out twice when Mission Local registered a test account.
Many programmers said the website needed to be taken down and fully rebuilt.
Forrest agreed, saying that it’s almost impossible to fix a product completely once it’s out in the wild. He said that in the past he’s had to weigh the balance between missing a deadline, and a product not working properly.
Jack Al-Kahwati, CEO of Velo Labs, had a different idea. You can have your 50 contractors hunting for and trying to keep the exchanges running, but you should also have at least one small, independent team of experts rebuilding the whole system from the ground up.
“Pull it back – nothing ever gets fixed on the fly. In my experience, you need to rebuild the whole thing, band-aid solutions never work,” said Karl Guttag, a 59-year-old hardware designer with a resume stretching all the way back to Texas Instruments in the 70’s. His personal website leads with an appropriate quote from a former coworker “If you haven’t tested it, then it doesn’t work.”