By the time you read this, we will have been inundated with the failings of the online enrollment system that was cobbled together in the service of the Affordable Care Act.
The politics are flying, and the first tiny bits of fact are starting to reveal themselves. The news outlets will do their best to create bite-size nuggets of dumbed-down bullet points. I shudder to anticipate the undulating day-glow colored charts, graphs and animations about to be unleashed on the American public to explain what really just comes down to common sense.
As you head into next year with excitement about investing in programs and systems that will take your agency to the next level, keep these high-level guidelines in mind. Repeat them into the mirror each morning. Make your technology leaders repeat them, and make your vendors repeat them.
If the administration had followed these rules, the site would have worked on day one, and I would have been forced to write this column about my cat. The Adventures of Mr. Bigglesworth was going to be the highlight of the year. Thank you, Mr. President.
1: Do not underestimate
Most technology projects struggle to even get off the ground. Why? As agency leaders, you know what you know, and you know what you don’t know. The gray areas carry a tremendous amount of risk, especially if you don’t have a trusted advisor at your side. Most agencies do have a trusted technology leader, but at the end of the day the person actually writing code will be someone who probably has no idea what the big picture looks like any more than you know how to optimize a recursive algorithm to find the probability of DataPoint Q at position X related to event T. This is the nature of software development, and the most important factor in your success is understanding just how hyper-complex it can be.
Imagine your goal is to get the entire general session of the United Nations to sing Sir Mix-a-Lot’s classic ballad “Baby Got Back” in three-part harmony while the translators are all on strike. While the language barrier is the most obvious concern, the complete lack of context, understanding and coordination are what will really sink your venture.
Did the administration underestimate the gravity of building healthcare.gov? I think the answer is pretty obvious. Everyone loves to point to Amazon as a model for a successful high-traffic marketplace, but it’s easy to forget that it’s been under constant development since 1994. That’s nearly 20 years of mistakes and fixes.
To add insult to injury, the government’s overconfidence led to an environment in which the non-Web options to enroll were barely even mentioned. This shifted everyone’s focus to the one method of delivery that had the greatest probability of failure.
2: Don’t believe high prices increase the likelihood of success
The cost of healthcare.gov is currently estimated at $133 million and counting. When faced with a complex problem and the realization that “smart people” are required to make it happen, the first instinct is to throw money at the problem. This strategy relies on a fatal binary logic flaw: Smart people are expensive; therefore expensive people are smart. This belief leads to a feeling of comfort based on the high price tag. It’s a nice thought. Unfortunately, it’s not true.
The administration just knew healthcare.gov would work. Why? Because it’s expensive, they told us, and that means we’ve hired the best. Then, when it failed publicly, everyone involved with the project demanded an immediate search for the “best and brightest,” and the fatal process started all over again.
A side note: The companies that secure government contracts are the companies that are good at navigating the government contract system, not necessarily the best companies for the job. Healthcare.gov isn’t a military drone, a construction project or a wrench. It’s a massive systems integration project, a merchant site, a data validation system and a workflow manager. So whom did the administration hire? General Dynamics. Northrop Grumman. CGI Federal. You’re going to throw these companies on top of a big pile of other companies and they’re going to create an Amazon.com in less than a year? Seriously? Does this sound like an achievable goal?
Spoiler alert: It’s not.
3: Keep it simple
I’d call this one a cliché except nobody ever listens to it.
When you hire 55 contractors who build 55 components of a system that interact with scores of other systems (built and maintained by countless other contractors), your project will fail. If you don’t test a system every step of the way, it will fail. If you take your vendors at their word and pass a rosy outlook on to your boss, Congress or the American people, you will fail.
Architecturally, there are just too many moving parts in version one of healthcare.gov. There’s a golden rule in software development: Don’t pack it all into version one. Attack one problem, do it well, make it elegant, launch it, then move on to the next problem. Now make it better without breaking the existing pieces. Use the launched system to guide the development of future features. The first iPod played music that you loaded from your computer (but only if you owned a Mac). Today it is a computer. If Apple tried to pack everything into the first iPod back in 2001, it would be BlackBerry today.
So what is the first feature that should have been the focus of healthcare.gov 1.0? How about showing Americans the available health plans? As of this writing, most people still have no idea what’s even available on the exchanges. This critical feature somehow fell to the bottom of the design list. If Amazon required you to create an account and provide a full family history before you could browse the available wool socks, you’d probably buy your wool socks elsewhere. If it didn’t work even after you acquiesced to the backward process, you’d never return.
In the real world, the gathering of your personal information happens at checkout, not on the homepage.
The last time I needed an IRS form I searched for it from the homepage on irs.gov and had it downloaded in the next click. The IRS website isn’t exactly the epitome of Web design, but it works like a charm. How do we get some of that action?
Somewhere along the line some bloated committee mixed together the needs of insurance carriers, government agencies and other ancillary parties and used this soup as a roadmap to build healthcare.gov. As usual, nobody thought to include the needs of the customer. The result? Epic fail.
As we head into 2014, let’s learn from these mistakes. Don’t run your business like the government runs its.