An IT Curmudgeon Looks at

Posted October 22nd, 2013 by Jim Ettwein

Ok, this is one of those rare non-partisan posts.  The events surrounding the unveiling and failures of Obamacare’s signature website stand out without political comment.  A bit of background: I’m an IT Management Consultant.  I have more than 35 years’ experience. I have implemented IT systems and websites in the US and around the world.  I have my own eCommerce website and know first-hand the issues surrounding such an implementation.

In 35 years there are, of course, some implemenations that were successful and some that had problems.  But I think that my success rate is well over 95%. And any problems were minor.

bad design’s implementation failed in so many areas it’s difficult to know where to start. So, let’s start at the top.

Funding. Clearly not a problem. Spending on this website has been greater than $300 million.  For those who use the internet many are familiar with high-volume websites that serve the public. Obama’s website is squarely in this category. Examples of successful implementations would include:,,, and There are even insurance websites that offer price comparisons similar to the health insurance exchange that’s part of Obamacare. is one example.

What do these sites have in common? They are very large, they handle huge unpredictable volumes, and they enable customers to comparison shop and purchase products and services online. So, it can be done. The major difference is that these sites are built as businesses, not as government websites. Businesses have a strong affinity to serving the customer in order to consummate the sale. The government knows it is the only game in town and isn’t necessarily worried about the customer.

So, what went wrong? It’s difficult to really know right now, since it’s doubtful that the facts behind the implementation of will become widely known.  What do I think went wrong? That’s easier to predict.

  1. Planning. Obamacare had a long time to prepare for the website. The law was enacted years ago, and planning should have begun at that time. We can go on the presumption that it did. However, it’s obvious that something went wrong in the planning since end-user testing (the customer) and transaction volume testing seemed to have either been left out (doubtful) or were insufficient to really test the site. Planning was insufficient and inadequate.
  2. User friendly design. Was the overall design friendly enough to please the customer and enable bad design 2them to accomplish the task: finding and buying insurance?  Clearly not. It’s difficult to get quotes, and not clear what final pricing might be.  In order to compare plans, one has to register first. It’s the process of registering in advance that many users will find concerning. Many customers are reluctant to provide personal information in advance when shopping for things online. This apprehension is stronger when the government is involved. Especially when the enforcement agency is the IRS. Many have a natural concern about providing information to the IRS that might draw extra attention to themselves. I do.
  3. Technology. Things that I’ve read indicate that current technology was not at the forefront of this site. Again, it’s not simple to know details, but current architecture should be able to handle spikes in use without any difficulty.
  4. Testing. This is one area that seems to be radically wrong. No site should ever go live without many levels of testing to insure its readiness. Simple low-level testing should indicate failures in basic modules. The next level is testing the entire application as a unit. Things like: Does it function from start to finish for a typical scenario? Is it understandable to the experience level of the typical user? Can uninitiated users navigate it easily? Can it handle a high volume of users? Can it handle spikes in usage?  Clearly there are severe issues in these areas. The typical user seems to not be the main focus as the Administration is hiring thousands of navigators to help users accomplish their task. High volume is also failing as the site cannot sustain a high-volume of users. I’d guess that someone failed in anticipating that there would be a huge volume during the first week when the spotlight would certainly be shining brightly on the site.
  5. Implementation. I’ve never been involved in an implementation where the developers were not standing by to try to fix last minute issues that always come up. In this case we have a comment from Secretary Sibelius that she’s asking the vendor to come in with high level people to try to fix this problem.  In fact, the government announced that they’re bring in Verizon to fix the problems. One wonders why the vendor who built it is not on the hook for fixing it. This is poor contract management since we’ll be spending extra money to fix a problem that should be covered by some sort of warranty by the developer.

Could commercial companies have done better?  Of course. They have in the past and will continue to do so in the future. Government has difficulty understanding what people want and need and isn’t worried that the customer will go somewhere else.

Testing was clearly a major trouble spot for this website. It was ill-designed and poorly executed. In looking for information about the Canadian company that designed and built it, there were many examples of failed attempts at implementing other government sites in the past.

So, what could possibly go wrong.  Go the website.  You can see for yourself.

Beware: Political Statement Ahead. Now, do you want to trust your healthcare to the same government who can’t even design and test a simple website? I don’t.

5 Responses to “An IT Curmudgeon Looks at”

  1. Varvara

    So, what went wrong? Let’s start at the top; the designers used 10 year old software that they ‘borrowed’. Testing; did they do any?

  2. Flick

    Excellent expose rr!

  3. Mark

    —–> jizya <—-
    Look it up. It will explain all you need to know.

  4. Casey Chapman

    They used a Canadian Company instead of an American one. GEICO stands for Government Employee Insurance Company. It seems to me, that it would have been a simple thing to call Geico, and ask them who they used. Of course, that would make too much sense.

  5. Chad

    I can’t believe they outsourced this job to Canada, of all places. Aren’t there any reputable web design firms in the US? Oh, but they probably didn’t support Obama or his healthcare law, so he would never use them.