I was doing research the other day and came across an interesting article on FederalTimes.com about an FBI information technology (IT) project that was estimated to cost approximately $451 million. When the article was written, in October 2010, the project was a year and a half behind schedule and $100 million over budget. Earlier in 2010, the FBI decided to use an "Agile methodology" to turn the project around and deliver by September 2011. Presumably, the FBI had been using a more traditional approach to this project, with a big up-front design.
The article sparked my interest in other government projects that cost taxpayers millions and billions of dollars. I found several billion-dollar projects, such as the Interior Department's $7-plus billion effort to consolidate the department's IT infrastructure across the agency, the Treasury Department's $2.8 billion IT project for telecommunications systems and support, and the FBI's $3.4 billion next-generation indentification system. In the Veterans Affairs Department, the financial-management system development program was canceled several times in 2004 and again in 2010 for failure to deliver results, despite a $250 million development cost.
As a taxpayer and an Agile enthusiast, I found my eyebrows going up as I read some of these estimates. I thought of surveys and reports I've read in the past that illustrated the success and failure rates of IT projects. One of my favorites was the 2004 CHAOS report by the Standish Group.
With the known statistical failure rates of projects, what is the cost of failure to taxpayers? According to the Standish Group's 2009 CHAOS report, the overall failure rate of projects has increased to 24%, from 2004's rate of 18%. The 2009 report also states that, during development, the rate of projects being challenged (for being late, over budget, not delivering desired functionality) is 44%, which is down from the 2004 report rate of 53%. According to the 2004 CHAOS report, the trends improved between 1994 and 2004 for project failure rates, but the rate of projects running over budget and over schedule is nearly flat at around 53%. From 2004 to 2009, the failure rate actually got worse, while the rate of projects being challenged improved marginally.
What do studies like the CHAOS reports tell us? That we are still risking a large percentage of our tax dollars on projects that have a high rate of failure or of running over budget and/or off schedule.
Projects are still too big in many cases, delivering functionality that never gets used and missing features that are really needed. It isn't fair to say all project failures or overruns are related to Waterfall, but certainly many miss the requirements, miss the real needs, or miss the business opportunities. Even though we cannot blame Waterfall, it does sound like these failures are endemic of Waterfall projects. The prioritization of features with Agile would reduce the risk of not delivering what's needed, when it's needed.
I did some "napkin" math and realized how great a risk the government is taking with my tax money. According to Jeff Zients, the deputy director for Management and Chief Performance Officer for the Office of Management and Budget (OMB), the federal government spends approximately $80 billion a year on IT projects. If the average failure rate is 24%, that means $19 billion a year is wasted. With a challenged project rate at 44%, nearly $35 billion a year is risked on projects that fail to meet their budgets or schedules.
The what-if scenarios of a WikiLeaks-type report come to mind. What if WikiLeaks put out information about the government's knowledge of project failure rates relating to Waterfall projects implementing big up-front designs, unrealistic schedules, and creeping scope while delivering functionality that never gets used? What if WikiLeaks told us that the government knew all along that projects using this methodology had only a 32% chance of being totally successful, meaning delivering what is needed on budget and on time? What if WikiLeaks revealed that 44% of the billions of dollars in spending ($35 billion) was being risked and 24% of the money ($19 billion) was being thrown away each year because projects failed to deliver? Would there be an Agile revolution? Would government agents start calling Waterfall proponents to forewarn them about the impending leak and urging them to take protective cover from the onslaught of critics waiting in the darkness?
We taxpayers usually want to be assured that the government is using our hard-earned tax dollars as wisely as possible. Yes, we know that there is bureaucracy in government and that there are programs that don't always make sense. But most people aren't aware of the cost of government IT projects. And until you dissect the numbers of the current annual budget for such projects, $80 billion might not seem like an unusually high amount of money spent on IT programs when you compare that cost to other expensive programs, such as Medicare, national defense, Social Security, and so on.
The waste that I am talking about is not unique to government; it happens in the private sector as well. It's just that governments are usually slower to change. Switching from Waterfall to Agile may not increase or decrease the government's projects budget of $80 billion per year, but it should increase the quality and performance of those projects. Much of the annual $80 billion goes directly to employing people to do the work. In an Agile environment, government staff and contractors might spend their time better and deliver products that more closely meet the needs of their customers.
It takes a lot to turn a government agency's methodology and process from a Waterfall environment to an Agile one. I have seen that firsthand as a consultant at a state agency. A government's environment presents many challenges and processes, and a history that doesn't make it easy to support Agile transformation.
But change can happen. There is movement in the Agile direction, and I see evidence that it's happening, slowly but surely, in our government. Take the FBI: It didn't continue with a methodology that wasn't working. Instead, those in charge made the decision to try a new approach to save the project -- and save millions of dollars. Another example among large federal agencies, according to USAspending.gov, is the U.S. Patent and Trademark Office, which is using Agile to develop a $45.5 million project called USPTO Patents End-to-End: Software Engineering (PE2E-SE). According to Accurev.com, the Department of Defense's Chief Management Officer Elizabeth McGrath announced the adoption of several Agile methodologies to improve the efficiency of development projects. And at the New York State Department of Labor, the CIO performed a wholesale change of the developent lifecycle from Waterfall to Agile, specifically using scrum.
President Obama is reevaluating IT spending and has proposed a 25-point plan to reform federal IT management. This plan will make programs more responsive to changing needs, echoing the fourth value in the Agile Manifesto, "Responding to change over following a plan."
The CIO of the Office of Management and Budget, Vivek Kundra, recently sent a memo titled "Memorandum for Heads of Executive Departments and Agencies" [PDF]. It calls on agency leaders to improve plans for high-priority projects. The first line of the memo states, "Federal information technology (IT) projects too often cost more than they should, take longer than necessary to deploy, and deliver solutions that do not meet our business needs." Thirty high-priority projects are identified on the USAspending.gov website and are listed here. The first line of Kundra's memo summarizes many of the feelings and issues that led to the 2001 meeting of 17 people at the Snowbird Ski Resort in Utah, where they created the Agile Manifesto.
I'm not sure we need something like WikiLeaks to tell us the government doesn't always spend our money wisely, or that there is waste. I don't think it would reveal to us anything we don't already know about Waterfall and Agile. So if we know there's a better way to do development projects, and our government knows it (according to my fictitious WikiLeaks proclamation), why is the government still risking 44% of the $80 billion a year budget and wasting an additional 24%?
As Scrum and Agile practitioners, perhaps there are actions we can take to effect a change in the way our nation does IT project work. Change in government is costly, but we ultimately want to either save money or get more for our money. As Agile citizens, I think we can do a few things:
(1) Become more aware of our IT spending within the government
(2) Lend a hand to those running projects
(3) Inform the government
(4) Keep quantifying results to help bolster Agile adoption and transformation
Don't wait for "AgiLeaks" to tell us what we already know. The time might be ripe for big change in the government's approach to IT management and delivery. I think President Obama's reform policy opens the door and offers an opportunity to inject Agile into the improvement process.