I’m going to go off script this post and suggest something provocative: Government is risk-averse, and the cost of risk avoidance is very high. Then I’ll recommend a way out.
First, I hope that it is not controversial to suggest that IT management in government is risk-averse.
This is not meant to be a personal admonishment but a recognition that, in government, your career can be hurt by a failure more than it will be helped by a success. In this environment, risk-taking is not the optimal path to get ahead.
Here is a real-life story about a telco, a baby Bell, when it was so bureaucratic and nearly a government unto itself. I ran a consulting firm that built data warehouses. We spun from Teradata in the early days and had a very high level of expertise in a new market. The baby Bell in question wanted to build a new data warehouse and invited us into the game late. My salesperson and I listened to the requirements and stayed up all night to develop a response for the next day. We built a plan from the bottom up, came to a reasonable estimate, doubled it to account for the risk of a fixed bid (they wanted us to take the chance since they were risk-averse) and then doubled it again to account for what we did not know after only 24 hours.
We waited to hear back, confident that we had enough room to hit the mark easily but unsure how our four-times estimate would stand up to the competition.
The next day, the executive in charge called us in and said, “Your proposal is one-third the cost of the proposal we received from our in-house, Big Five, consultancy. What is more, your proposal demonstrates such a high degree of competence that we are convinced that you could actually deliver for that amount.” We were pumped! He then went on to say, “But my boss, a senior executive, is unwilling to take the risk. He believes that if you fail, even if it is a small chance, that we all will be in trouble. If our Big Five consultants fail, they will be blamed, and we will be absolved.”
So they agreed to a deal that was twelve times our original estimate. The Big Five overran its estimate by two times, a 24-times difference, and the project failed even after that.
In my last post, I suggested that there is no such thing as a $200 million software startup. The very big startups spend $20 million to $40 million on IT to get a minimum viable product. Most spend well less than that.
So why does the government regularly commit to $200 million programs? The answer is risk aversion.
Take a more than $10 million project and double it and double it again and then convince yourself that only a big firm can handle a $40 million-plus project and you will end up with an $80 million to $100 million bid over five years.
Please do not view this as a cheap shot at the big consulting companies. They are risk-averse, more or less. They are used to competing against each other and having the same cost structures. The issue is that the playing field has not changed for 30 years.
One way to reduce these costs is to work on a time and materials (T&M) basis. Stop putting out fixed-bid contracts or capped contracts where the risk shifts to the contractor. When you shift the risk, you pay for it.
With a T&M contract, you get what you pay for. If you combine T&M with Agile methods, you should see what you are paying for after every two-week sprint. If the requirements change and new user stories are inserted and prioritized, then you just adjust. There is no contentious change management process to sap everyone’s time and energy.
To be sure, with T&M, the agency has the responsibility for managing the costs and measuring them against the benefits in short increments. An Agile approach supports this, but you have to learn to manage agile projects.
There is no magic to Agile methods. The reason that you can cut your IT development costs with proper Agile project management is that you manage the risk rather than paying several multipliers of the value to have it handled for you. With Agile, you can run your projects like a Silicon Valley company, like a digital service company, and eliminate the doubling and doubling and doubling again of the costs.
Further, you can do this with your favorite big consulting partner, and they will be thrilled. They would prefer a steady profit to the risk any day.
To tie this back to the beginning: If government IT could take some risk and learn to manage it using the same Agile methods used by the best commercial development houses, IT costs will drop and the number of successes per taxpayer dollar will go way up.
Rob Klopp is part of the GovLoop Featured Contributor program, where we feature articles by government voices from all across the country (and world!). To see more Featured Contributor posts, click here.