Maybe you should re-think that big IT transformation effort

Steven Kelman
5 min readOct 2, 2019


I have known John Kost since 1992, just before he became the state of Michigan’s first CIO and he was participating in a Kennedy School program directed by Jerry Mechling called “Strategic computing in the public sector.” (Mechling was fairly early out of the gate in promoting the idea that computing should not be seen mostly as a back-office function, but rather be tied to mission; he is a four-time Fed 100 winner.) Since 2002 Kost has worked at Gartner, where his title is currently Distinguished Vice President.

I just recently got an email from him, occasioned by his seeing an article in the Kennedy School alumni magazine on failed digital government projects. (I co-teach a course called Averting Digital Disaster that centers on In the email, Kost enclosed a Gartner report he had co-written with Rick Howard called “Where the Buck Really Stops for Government IT Project Failure.” (The full report is behind a Gartner paywall.)

The report has a very dramatic theme and recommendation. If the project is mission-critical and will require significant changes in agency business processes, the buck really stops, the report opines, with senior government executives, at the agency head or deputy level. “Much of the cause of project failure is actually unrelated to the IT department itself or to the technology,” Kost and Howard write. “More often, it is the unwillingness or inability of senior government executives to engage in effective decision making.”

But the report goes much further. “An executive business sponsor must be personally invested and visibly committed to the success of any business initiative that involves significant organizational and process changes produced by the adoption of new IT solutions or services.”

This advice obviously does not apply to all or even most IT projects — only to those few that are both mission-critical and require significant changes within the agency.

Before a government organization undertakes a mission-critical project with an IT component, the report states, “the CIO should assess the aptitude, willingness and readiness of the organization’s executive leadership to do what’s necessary to make the project succeed.” Then the authors add the kicker: “If the organization is not in that position, consider non-IT alternatives for achieving the policy goals.”

In other words, without senior executive participate on, you should probably scrap the project before you start. “There are no caveats here,” Kost and Howard write. “There are no workarounds. There are no ways to outsource this fact.”

Why? A key reason, the report argues, is that senior executive involvement is required for the change management efforts that will always be necessary for a project requiring organizational and process changes to have a chance. “For major programs and projects in which the intended objective and result will be significant changes in business processes, it is only the senior leadership of the organization that has the clout to enforce these changes across the organization. So, unless issues of change management have close attention paid to them, the IT embedded in the project may be successful, but the business changes will never get implemented and the results of all of the effort expended becomes a waste of time.”

The following steps, the authors argue, need to be worked out before the project is started:

  1. The executive business sponsor or sponsors must know how much time is required, and have the time and ability to provide the necessary level of personal engagement in the project (emphasis in original).
  2. The executive business sponsors must perceive this project as a business project, rather than an “IT project,” because they understand it affects the mission of the agency.
  3. It must be clear who the decision makers are, and the path for policy decisions (governance) must clear and timely.
  4. It must be clear who will resolve conflicts whenever the inevitable policy, funding or contractual issues arise, and the process for escalating those conflicts must be well defined.

During execution, the “project manager reports directly to” the agency head or deputy, and the executive leader has “constant and direct interaction with direct reports on execution of change management.”

The report contrasts these steps with what they call “typical” for top executive engagement, even for these kinds of mission-critical projects.After an “obligatory sign-off” by the agency head on a business case, there are “perfunctory quarterly steering committee meetings.” Change management is “delegated to project manager.”

These projects are difficult. The report states: “We will often ask for an audience with the [agency head] to inquire simply, ‘Do you know what you are getting yourself into?’”

This advice is really tough love for government executives. It is a credit to Gartner — since consultants often don’t want to offend the client — that the report says outright that if top executives are not involved, an effort that relies significantly on IT to realize the agency’s goals probably should be cancelled before it starts.

Reading Kost’s report, I discovered that it is hardly hot off the presses — it came out five years ago. I spoke with him about the work, and he acknowledged that it hasn’t taken in the way he’d like. In the federal government, agency heads or deputies typically do not spend quality time on IT, probably have never heard of Gartner (which deals mostly with the IT shop), and think that personal involvement in IT projects is “beneath them.” However, most state CIOs report directly to the governor, and this has helped Gartner, through those CIOs, get executive access to pitch this approach.

And in Australia, where the state government in Queensland fell following a high-profile IT failure, the head of the civil service invited Kost in to discuss his ideas, and since then there has been some movement there and in some other Australian states.

The perpetual optimist in me notes that even in the last five years the visibility and centrality of digital services in government have increased dramatically. Today, making the case for top-official involvement might be somewhat easier. But for the foreseeable future the digital evangelists in Washington are likely to be mainly CIOs and some good government groups.

So don’t hold your breath for a snowball building. But let’s hope some cabinet secretary, deputy secretary or assistant secretary chooses to take up this cudgel and get things started.

Posted by Steve Kelman on Oct 02, 2019 at 2:13 PM