I still remember playing Little League as a kid - it took a home run or at least 2 hits to get your name in the Daily Gazette, our local newspaper. As a 12 year-old kid in upstate NY it was always a thrill to wake up in the morning and see your name in print. Recently, the Social Security Administration made the paper for striking out, and this is one headline they certainly weren't looking forward to.
The ABC News story certainly cherry picked a few catchy quotes that were written to grab headlines. It's easy to pile-on when an organization spends almost $300M on a system that still doesn't work. However, look a little closer at the McKinsey report that the story references and you’ll get a much better picture of what really happened. My overall assessment is that the 75-page report is fair, thorough, and a ‘must read’ for any and every organization considering a dynamic case management initiative.
When writing an article or blog you never want to bury the lede. In this case I'm not even sure what the lede is. Between the McKinsey report, the technical decisions that were made, the software that was chosen, and the $300M, I have enough material to blog for the rest of the year. For this article let's just focus on a few of the main blunders.
The McKinsey report, while specific to the DCPS initiative, is applicable to any enterprise dynamic case management initiative. The report is an easy read and will make sense to any audience - from the executive to the developer. It clearly outlines several challenges facing DCPS and potential areas for improvement, most of which revolve around adopting Agile software development methodologies and improved program management.
Many of the McKinsey suggestions for SSA are practices that we take for granted at srcLogic. To us, they are the gold standard. Our methodology is based on principles and best practices from Agile, Lean, and the GE Change Acceleration Process (CAP). Most people are familiar with Agile & Lean, but the GE CAP Model is less known. Those experienced with the GE CAP model wouldn't be surprised to read the beginning of the McKinsey report which acknowledges that SSA had a sound plan and foundation for success. This is one of the core tenets of the GE CAP: most failed projects had sound technical plans.
So if SSA had a great plan, what went wrong?
Here is my Cliff Notes version of the McKinsey recommendations for SSA and DCPS:
- Select a single executive leader with end-to-end responsibility for delivering DCPS.
I included this in the list because it’s a no-brainer. In fact, I am astounded that SSA was allowed to spend $300M without a single executive in charge. How is that even possible? When I started consulting this was one of the first edicts that we learned during training. The theory being that multiple people can’t share responsibility for something. If multiple people are responsible than no one is actually responsible.
- Adopt additional Agile practices to strengthen alignment with business:
- Implement select Agile tools including: single repository for prioritizing requirements, backlog, visual board with burndown chart, etc
- Develop in Agile Sprints with frequent checkpoints to de-risk releases and provide responsiveness to user priorities
- Create product owner role for each DCPS Component
Again, I am always astonished when I see organizations that develop with a waterfall approach. The one constant I see on every project is that end-users don’t know what they want until they see it. Once you accept that basic premise, you will be exponentially more successful. Nothing is worse than a user acceptance test that turns into a requirements session – and believe me this is a common occurrence on waterfall based projects. The key to Agile is engaging the SME’s and end-users throughout the full lifecycle, not simply once at the beginning and end of the process.
- Evaluate “next best alternative” including commercial-off-the-shelf (COTS) options in terms of functionality, architectural flexibility, cost, schedule, and risk
Full disclosure, I’m not 100% certain which specific tools and technologies were used on this project. I did do some digging around and have a pretty good guess, though I won’t post it here. I will say that I frequently see this phenomenon on integrator led projects across the federal government. I call it the “too clever by half phenomenon”, where proven industry leading tools are eschewed in favor of either custom development or the open-source Frankenstack. When you are looking at enterprise rules-driven dynamic case management your short list should have one name on it: Pega. If you outright reject that solution, there are 2 or 3 more that could work, but be prepared to validate why you’re forgoing the industry's leading platform. If you have a $300M budget, cost is not one of them.
Beyond the McKinsey recommendations there is one more point that I would personally add: Replacing an existing system is significantly more complicated than building a new capability that doesn't currently exist.
This is a topic worthy of its own blog but something that I am surprised McKinsey didn't mention.
So what now?
At this point the $300M is spent. The best we can do is to view this as an absurdly expensive learning experience. As a taxpayer I hope SSA is able to salvage the investment or cut the cord. As a technologist, BPM Software practitioner, and dynamic case management evangelist I hope other organizations are able to learn from the SSA’s mistakes and adopt the ounce of prevention rather than the pound of cure.