Buying insurance on the new health exchange should be as easy as booking a flight—both have your life at stake. Yet where the airline industry has arguably perfected this customer experience online, the health exchanges that went live last week all but ignore years of learning by UX practitioners. Todd Park, an advocate of applying the Agile method in building platforms[1] and U.S. Chief Technology Officer, said in USA Today, "These bugs were functions of volume […] Take away the volume and it works.”[2] In other words, blame the user, not the developer.

Park is severely mistaken here and is underlining the issue with the hubris of developers who seek lean approaches to large-scale projects. The Affordable Care Act (“Obamacare”) Health Exchange is a large-scale, high-stakes, high-profile initiative. The goal is not to build some start-up MVP (minimal viable product) where, like the technology company Uber, a pilot in one market is tested with the easiest usable interface. The way I see it, our government fell for the allure of Lean UX when it needed anything but. In going through the exchange myself, I found simple key errors in the overall user experience. As a UX practitioner who empathizes with people less technically literate, I found the experience trying, frustrating, and disappointing. (Updated: Many Lean UX fanboys on Twitter have argued that there's no proof that there was lean applied here, my assertion is purely subjective based on circumstantial evidence. Hence why this blog post is on my website.)

Remember at the beginning of AOL’s subscription service demise in 1997, when demand for the dial-up caused busy signals for hours? (Ironically, a problem rooted at AOL’s headquarters in the Washington, DC suburb of Vienna.) You would think those early-day problems would be antecedents in making sure you had more than enough bandwidth to meet demand and then scale down. Yet, under Park’s leadership, the government went lean with scalability in mind as a reactive measure rather than a proactive SDLC (software development lifecycle) with a proven, risk-averse approach. Rather than go through a scoping phase to tick off criteria that assess whether to go lean or not, from what I can tell in experiencing the website experience, Park & Co. went with a one-size-fits-all approach for User X instead of Users X,Y, and Z. It’s doubtful they went through personas as the design would indicate.

(Updated: I think it's fair to say that a Lean commandment is to go with the lowest minimal elements to meet the need and scale up to users. I argue that the number of servers you have available on a launch date is an aspect of the UX. Arguably, saying you planned for 50,000 visitors at once, is Lean thinking.

I'm a fan of Lean UX when you have a start-up or a small team that can be nimble and iterate. When dealing with large, global initiatives or big organizations that serve a diverse user base, private or public sector, I've yet to experience positive meaningful outcomes. I think Lean UX needs to evolve with a checklist to see if Lean is really the right way to go when planning a project, weighing it against the desired outcomes or knowing when the Lean approach may not be sustainable, something I cover in my Muscle UX book I'm writing now. Surely, we cannot approach everything by a one-method-fits-all myopia.)

Now, let's take a look at the homepage for healthcare.gov (snapshot on October 7):

1) Why not use self-segmentation? Why overwhelm a user with so many choices and the burden of having to sort out where to go? Imagine that you’ve never been on a computer, you’re sitting with a caregiver trying to figure this out, and have just been shown a bunch of directions to go in. We know from Columbia’s “Jams Study” that users are more apt to engage and have less confusion and frustration by having choices limited[3]. Like many pharmaceutical and health care sites, there should be a model window immediately prompting the user to choose between being an individual or family, or employer. Take a look at how Oscar Insurance does this elegantly (with an accessible, large font).

2) Log-in not accompanied by a “register” or “sign-up”? With such a complicated subject matter, going with what users know and expect is probably best. In fact, when President Clinton gave his pitch to the country during a chat with President Obama, he urged young people to get enrolled. Yet, studies show that millennials surf with temporal expectation bias (e.g. the search bar is perceived as always being on the upper right)[4].

3) Apply Now button has accessibility issues and is fighting to stand out. Someone with deuteranopia, colorblindness, would see this button as white in yellow with a very challenging contrast. Also, the button is designed to blend in with the picture in the background; hardly the standard 70% contrast level for a call-to-action. Making the button larger and starker makes this processing easier. Also, in the eye-tracking studies I’ve done, users almost always connect with a face that is breaking the “fourth wall” (looking right at you). (Updated: Should it be "Apply" or should it be "Enroll Now" given that the context to "apply" is that a user may feel like they could get rejected like credit cards or college admission. According to this report, 99.6% of healthcare.gov visitors do not enroll!) While the apply button is on the left, the face will be the first thing a user sees and connects with, so put this face on the left and the apply button right next to it. This button also competes with…

4) Start Here—this call-to-action button is not at the required 70% contrast level, and if I’m a user, which is it? Am I to Start Here or Apply Now?

5) Washed out FAQ section is hardly distinguishable within this design. What we have here is design without UX involvement. While, yes, this is visually pleasing, there is a gestalt effect blending these important FAQs, causing them to get lost. This site would be better off having a lot more white space with higher contrast, which would make it very usable for anyone visiting. Not to mention, if you look at the responsive site in mobile you’ll see an easier to read vertical list. (Easier because reading long horizontal menus takes users more time).

6) Urgency is buried. Tiny numbers appear beside even tinier print with extremely important information. To give the sense of urgency (and also a sense of confidence that users have time to sign up), these important numbers need to be pushed up to the top and clearly enlarged. (Do they really need to have the “Open Enrollment Began” date?)

7) Blog looks dated already. Either have daily updates or don’t promote it. The site looks abandoned, not to mention that it doesn’t even address the elephant in the room: there may be some issues signing up.

8) Connect with Us? Why? Without an example of a story, why would a user want to share one? A nicely embedded video with a call to action to share would have a bigger impact.

9) I’m following why? Give context…will following help users stay up-to-date and informed?

If this is the UX assessment of just the homepage, you can imagine what the sign-up experience is like. In the hour it took me to get through the New York health exchange, led there by HealthCare.gov, I saw filters resetting once the search was made, causing me to lose all the filters I had set without a contextual summary; an alphabetical dropdown menu of “medal” plans rather than in hierarchy—Bronze, Gold, Platinum, Silver (a terrible nomenclature that really doesn’t invoke the best feeling about signing up and confuses metal with medal. I mean who wants to be bronze? Well, if you can’t afford it, that’s the medal you get, yay!); no easy way to distinguish differences between plans (they would do better to have a chart similar to the booking of first, business, and economy classes on an airline); being told after I tried to search by doctor that the feature was unavailable at this time; etc.

To be fair, I’m not entirely sure I can hold the U.S. accountable for individual state exchange experiences, but certainly I can blame Park and our Federal Information Technology team for not setting an exemplary example.

(Updated: After I posted this, @Usertesting posted this tweet, which boasts a usability test for the site, giving it a 2 out of 10, highlighting some of my UX points and conducting a deep dive in the usability via actual user testing):

Great @Obamacare usability study by @ctomlin http://t.co/22eelb7E1T. Great holistic UX review by @Griffopolis http://t.co/9eZ5s9Nzxe
— UserTesting (@usertesting) October 10, 2013

Denis Griffith (@griffopolis) had led global UX initiatives for Emirates Airline, launching iteratively this year, while XCD/head of UX at Atmosphere Proximity. Denis also served as director of UX with Havas Life, helping to create one of the first presentation builder and sales aid iPad apps. He is currently working on a book about Muscle UX due out in spring 2014.

[1] http://strata.oreilly.com/2012/05/us-cto-seeks-to-scale-agile-te.html

[2] http://www.theatlanticwire.com/politics/2013/10/obama-administration-opens-about-exchange-glitches/70248/

[3] http://www.columbia.edu/~ss957/articles/Choice_is_Demotivating.pdf

[4] http://users.wpi.edu/~djamasbi/PDFS/2010%20djamasbi%20et%20al-%20IJHCS%20(journal%20version).pdf