As I pen these words, the Obama administration is in the midst of an absolute debacle over the botched launch of Healthcare.gov. Regardless of what you think of Obamacare, it's hard to ignore what has become one of the most visible IT project failures of recent memory. Details are still coming out, but there's no shortage of opinions on what's wrong – and how to fix the website by December 15th.

In case you're unfamiliar with the particulars, a Time post details some of the technical problems with the website:

The basic architecture of the site, built by federal contractors overseen by the Department of Health and Human Services, was flawed in design, poorly tested, and ultimately not functional. “You need there to be good people on the inside to make good contracting decisions and good people on the outside to do the work,” explained Clay Johnson, a Democratic technology consultant who recently worked as a White House fellow. “Right now, it’s the blind leading the blind.”

Even on the back end of the site, data was garbled and, in some cases, unusable. The nightly reports that insurance companies receive from the federal government on new enrollees in the health plans have been riddled with errors, including syntax mistakes, and transposed or duplicate data, according to industry veterans. In other cases, insurers received multiple enrollments and cancelations from the same person, but since the documents lacked timestamps, it has been impossible to know which form is the most recent.

And here's where it gets downright embarrassing:

Companies have resorted to contacting enrollees directly to get answers, a solution possible only because so few have been able to sign up. "We are seeing and hearing that enrollment files going to carriers are incomplete, there are errors,” said Dan Schuyler, a director of exchange technology at Leavitt Partners, a firm that consulted with several states in setting up their websites. “In three weeks or so when they start receiving these in mass volume, tens of thousands per day, it doesn’t matter if there’s a 1 percent error rate. Insurers don’t have resources to go through them and clean them up.”

We've seen plenty of these project failures before, although arguably never one with such a public spotlight on it. (This was not a small business implementing a new CRM application.) Obama made healthcare reform one of the central pieces of his presidency, spending enormous political capital in the process.

Annoyances like poor system performance are one thing, especially with such buildup to the site's launch. Inaccurate data on something as critical as health insurance is quite another. While reading many of the analyses and critiques of Healthcare.gov, I was reminded of something I like to call the data equation:

bad design + no testing --> bad data

Simon says

What's done is done. In an odd way, the abject failure of Healthcare.gov is actually beneficial – at least compared to a "partial" success. What if "only" ten percent of the people who tried to use the site could not? What if "only" ten percent of the data was suspect? It's conceivable that the administration wouldn't have pulled the plug. While some people would have successfully signed up for coverage, hundreds of thousands or perhaps millions of others would have been thwarted. Can you imagine the lawsuits and bad PR?

When building new systems and websites, unrealistic deadlines, vendor competition and a lack of testing often yield duplicate and incomplete data. Learn from the mistakes of Healthcare.gov.

Feedback

What say you?