If it can happen to Obama, it can happen to anyone (yes, including you)!
The much publicised “Obamacare” or “Affordable Care Act Insurance” is certainly a solid example of a technology related catastrophic event in recent times.
I am certainly not doubting the intention of the website. Giving well needed insurance to those who could not afford it, goes back to President Clintons promise of health care reform in the early 1990’s. The intention was to help those in need.
Initial alarms seem to have pointed to serious performance issues with the website rollout.
What are performance issues? This is something when a website or piece of software runs slowly when there are a high amount of customers logging on at the same time. A symptom of this can be response times of 10, 15, 20 or more seconds when using the site. Worst of all it can crash altogether. This is what occurred consistently with the ObamaCare website.
Claims are that the website was failing under test conditions with 500 users logging into the system. The system design and infrastructure had to handle several million US citizens requiring affordable insurance.
The website continually experienced error messages and crashes upon launch. Very much a symptom of what occurred above. We can only speculate, but the software might have never been in a position to be adequately load tested due to functional defects perhaps. Or the IT vendor(s) employed to undertake the work possessed a low quality test capability.
Fingers were squarely pointed at the Obama management running his “Affordable Care Act Insurance” program and the associated management of the vendors building the website.
CMS spokesperson said they under-estimated the amount of users that would use the system.
“How on earth they under-estimated that got me stumped!”
Appears communication had broken down between the program and Obamas statistician advisers perhaps. Not uncommon to be honest as this happens in many businesses with sales/marketing representatives not communicating adequately with technology representatives regarding sales projections.
Quite simply, the site should had never been launched and should have been delayed. These are the hard decisions that must be made with high profile software. The right and accurate information must be available to make informed decisions. In this case I question the information channels to Obama or trusted advisers.
I strongly believe that incorrect or bent truth was sent to Obama leading up to launch, but we must not also forget the passion and the persuasive abilities of a leader such as Barack Obama. Health insurance reform was a policy he desperately wanted to succeed. Maybe his direct reports were too scared to tell him the truth?
I have been in situations where a passionate business sponsor convinced everyone that an idea would be successful without an upfront ROI assessment. Funding was granted, software was developed and then the software was pulled upon launch. Basically because someone spoke up at the last minute and challenged the profitability of the idea. A similar situation could have occurred in Obama’s administration.
Either way you look at it, Obama had to address plenty of public scrutiny during high pressure press conferences.
Below are 2 short extracts from US media sources of Obama answering questions post launch of the website. They go straight to the root cause of the problems outlined in this blog to date.
1/ Obama on : ‘No Excuse’ for Obamacare Website Failure Obama Vows Fixes To Health Insurance Website — https://www.youtube.com/watch?v=nZZOx1k_t7A
2/ Obama not informed on website failure — https://www.youtube.com/watch?v=igsTxgNMyYI
While the opening speech praises the positive intention of the website, it doesn’t paint a great picture of the people developing and testing the software. Basically makes them look incompetent. The above 2 youtube clips are an excellent case study or human highlight film of a technology public relation disaster. (No matter how good your speech writers are…..)
To conclude, adding more server capacity doesn’t always solve a performance problem. Sometimes the code that sits on the server/hardware needs to be optimised to handle high volumes of users. In this case the website launch had 250,000 users access the site instantly. Ironically testing had not successfully reached 10,000 concurrent users prior to launch. Oh dear!
From my experience, unqualified people make decisions to launch and experience the above type of public relations disasters. Make sure the right people are advising you when launching your next software project to market.
What have we learnt in this instance?
- The ObamaCare end to end software development and particularly the testing capability appeared to be extremely poor.
- Just because your software might work with one user, that doesn’t mean that will correlate when your entire customer base logs into it.
- Choose the correct IT capability to build your software.
- Your old long-term IT vendor might not be the right fit for your next project.
- Sometimes a leader’s passion for something can get in the way of achieving the right outcome.
I continually see customer volume simulation or performance testing overlooked when building or modifying software. Some business owners or non-technical stakeholders don’t understand it, which causes the continual oversight and general lack of funding to support it.
Know your risk profile and demand that of your IT vendors, management and reporting. Most importantly empower your software testing capability to expose risk that could impact your brand reputation.
- Performance testing doesn’t require expensive infrastructure 24/7. Sometimes it only requires it during peak periods. E.g. being a government associated website leading up to, during and after an election.
- There are cost effective cloud solutions in the market that allow easy performance testing that pay per the hour such as; Loadimpact.com, Blitz.io, Loadbooster.com (These types of companies could have easily supported the testing requirements for the ObamaCare website)