Over the last 18 months or so I have been following Vivek Kundra’s adventure as the US Federal CIO. I had the pleasure to meet Vivek when he was the CTO in Washington DC and we shared our views of web 2.0 and cloud in what happened to be the early days of a much broader and probably more rapid evolution than we would have thought. I also have a chance to meet him briefly every time I am in Washington, as he is always kind enough to make time for me.
I vividly remember our last meeting in October, when he told me that the second half of his tenure in the current administration would have been less about vision and more about execution. He actually answered my question about how he would make the most of a role which is inevitably influenced by the pace of political cycles, before I even asked.
However, being the cynical analyst that I am, I was not entirely sure how fast and how deep his turn toward implementation would be.
Well, I am very happy to see that he proved my cynicism unfounded by publishing 25 Point Implementation Plan to Reform Federal IT Management, which he and Jeff Zients – Federal Chief Performance Officer – presented on December 9 (see webinar replay).
While Gartner will publish an official note with analysis and advice to our clients in the coming days, I’d like to offer my personal view, also in view of my good personal relation with Vivek.
So here is the list of 25 points, and my personal take on each of them.
A. Apply “Light Technology”and Shared Solutions
1. Complete detailed implementation plans to consolidate at least 800 data centers by 2015. (good)
Here the plan suggests a combination of actions that individual agencies need to take, the establishment whole-of-government task force to review process and the implementation of a dashboard to share progress with the public at large. Indeed a smart move to apply multiple pressure points, knowing that different agencies at different stages of their consolidation progress will need different stimuli.
2. Create a government-wide marketplace for data center availability. (very good)
This is one of the coolest ideas of the whole lot and shows that Vivek and his colleagues at the Office of Management and Budget are truly shifting from vision to implementation. While there is still a considerable push for using cloud-based solutions, the message here is that if agency A runs out of computing capacity and agency B still has excess capacity, they can trade such capacity. An alternative would have been to look at a government cloud model with multiple agencies as providers, but the proposed marketplace approach is neutral with respect to the specific computing or sourcing model. Do you have a private cloud? Make it available on the marketplace. Do you have just rack space available? Make that known on the same marketplace. Brilliant, although not easy to implement.
3. Shift to a “Cloud First” policy (good but …)
Here it comes, and I like that Vivek keeps pushing for “commercial cloud” first, to then mention private cloud and finally community clouds (state or regional). This is in line with my observation that the relevance of “community” (or government) clouds is less than many originally thought. If vendors want to build government clouds as part of their commercial offering, so be it, but let’s avoid the proliferation of agency data centers turned into community cloud: the marketplace at point 2 solves that problem.
On the downside, while the announcement of a cloud strategy was overdue, the idea of requiring each agency to identify three moves to the cloud reminds me of the requirement of having three high value open data sets as part of the Open Government Directive: some agencies could address their cloud strategy in “compliance mode”, rather than using the various sourcing options delineated by the plan in the most effective way.
4. Stand-up contract vehicles for secure IaaS solutions (ok, but…)
This recalls the upcoming availability of IaaS services on apps.gov, after the recent award by GSA. However the plan sounds a bit optimistic on the FISMA Certification and Accreditation side. In fact, until FedRAMP does come to fruition, agencies still need to go individually through a C&A process with the vendors on the list. Also, the list of vendors does not include some of the big players in the cloud market, and it is unlikely that those selected by GSA will be the only ones.
5. Stand-up contract vehicles for commodity services (ok, but…)
This announces that what happened for IaaS could happen for cloud-based email, with GSA moving in about a year time toward government-wide contract vehicles. Given how intense the battle for cloud email already is (see here and here) I suspect GSA may be a bit late.
6. Develop a strategy for shared services (good)
It was about time to revisit the Lines-of-Business shared service strategy from the Bush administration. Vivek has publicly praised some of the LoB approach in the past and the plan seems to be willing to build on that experience, and improve service levels and sustainability of shared services.
B. Strengthen Program Management
7. Design a formal IT program management career path and 8. Scale IT program management career path government-wide (very good)
While in the first half of his tenure Vivek has focused on identifying shortcomings in program management, the new plan is to establish this as a recognized profession with its own career path. If program management is a problem, let’s create a context to attract and retain the best. Also approaching this with a pilot before deploying it more broadly is a smart idea.
9. Require integrated program teams (very good)
Recognizing that different roles and skills need to be involved and become individually and jointly accountable for program success is a milestone in solving some of the thorniest problems affecting IT programs. This may not be easy to implement, taking into account that people may move as a consequence of different career paths and dynamics, but definitely worth pursuing.
10. Launch a best practices collaboration platform (ok, but…)
I am a bit allergic to the term “Best Practices”, as this has been overused in government. I assume that what is meant here is “good practices” or simply “something that worked”. However what is essential is to be able to capture the details of the context that made a particular practice work, so as to avoid that people try to repeat somebody else’s success and then fail because there were differences they did not take into account. A collaboration portal is a good thing, but nothing replaces the value of spending time together. Of course technology can play its role there as well.
11. Launch technology fellows program (good)
Projects and programs at the federal level have scale and complexity that are unprecedented in almost any other industry sector. Creating mechanisms to make this context more accessible and exciting for bright technologists, and strengthening links with academia is an effective way to bridge the gap. On the other hand, there are other aspects of working in government that may put off technology geeks, such as ability to join and actively participate in social networks for both professional and personal reasons.
12. Enable IT program manager mobility across government and industry (good)
Making transition from industry to government and back much easier than it is today is essential to make sure that the best skills are available to government but also to benefit industry with the experience of professionals who have been confronted with hugely challenging programs in the federal government. Also greater job rotation inside government is great. However program success is tightly linked to understanding the context, and just looking (worldwide) at how long it takes to brilliant IT professionals and managers to get up to speed with government when they come from the private sector (and vice versa), reasonable expectations need to be set.
C1. Align the Acquisition Process with the Technology Cycle
13. Design and develop a cadre of specialized IT acquisition professionals (good)
IT is complex and markets keep changing, therefore acquisition professionals are constantly challenged. Making sure that people who are involved in increasingly complex IT acquisitions have – or have access to – the necessary skills is of paramount importance. I just wonder, if Vivek looks at some aspects of IT as a commodity, shouldn’t IT buyers be better at buying rather than at IT?
14. Identify IT acquisition best practices and adopt government-wide (good)
Once more, great idea, but let’s make sure that experience comes from beyond just IT and the context of what makes a practice “best” is duly captured.
15. Issue contracting guidance and templates to support modular development (very good)
One of the issues afflicting IT program is that the complexity of the acquisition process discourages modular development,although everybody would agree it is the lowest risk bath to program success. Bringing as many IT program managers and acquisition officers up to speed with how to encourage and even force modular development by adequately combining existing contracting vehicles is a big step ahead in moving this from a vendor to a buyer market.
16. Reduce barriers to entry for small innovative technology companies (ok, but…)
I am personally very much in favor of broadening access to government IT contracts to smaller companies. However the plan emphasizes only the barriers to entry from the vendor’s size, but does not mention how IT program managers and acquisition officers should deal with the level of risk that smaller companies inevitably bring to the table, often being financially more fragile or possible target of acquisition. Maybe this is something that points 13 to 15 above will cover, but IMHO buyers need to be better prepared at assessing the nature of risk associated to engaging smaller players.Incidentally, involving large players and incumbents is far from risk-free, as the many failures showed over the last several years: what I mean is that we are talking about different categories of risk, and government officials are less used at dealing with those associated to smaller or newer players.
C2. Align the Budget Process with the Technology Cycle
17. Work with Congress to develop IT budget models that align with
modular development (very good, but…)
Whereas the other points so far are about improvements or significant evolutions within the boundaries of existing regulations, this section deals with deeper changes that are absolutely essential to make most of the above work. Therefore Vivek’s ad OMB’s willingness to address Congress and agency leadership in order for budgets to be used more flexibly (as the plan says, “at the point of execution, rather than years in advance”) is laudable. It is an essential precondition to make IT portfolio management part of the normal course of business for all agencies, rather than a desirable option or a “best practice”.
However when it comes to changing the way money is used and giving agencies greater discretionary power, there might be all sort of political hurdles. On the upside, though, Vivek’s relentless fight for transparency may actually play in favor of change, as the usual criticism that giving more room for maneuver to agencies makes oversight more difficult, is answered by the increasing use of publicly accessible dashboards.
18. Develop supporting materials and guidance for flexible IT budget models (good)
Indeed it is not only a matter of convincing Congress, but also of making sure that Agency CIOs and other executives actually understand the options they have to use IT budgets more flexibly. This point also cover the greater transparency required and mentioned above.
19. Work with Congress to scale flexible IT budget models more broadly (very good)
Smart move to start small and scale with continuous buy-in by Congress. Let’s make sure they are comfortable with small examples, before moving to the next level.
20. Work with Congress to consolidate commodity IT spending under Agency CIO (very good)
Giving CIOs the ability to buy relatively straightforward stuff more easily and consistently is an important step toward rationalization and consolidation. I particularly like the fact that the plan does not advocate for top-down consolidation (i.e. “the OMB or GSA will chose for others”) but makes this a more organic process, mostly left to Agency CIOs (who often have significant consolidation challenges within their own agencies), before moving to the next level of consolidation.
D. Streamline Governance and Improve Accountability
21. Reform and strengthen Investment Review Boards (good, but…)
This section of the plan complements and pulls together most of the points in section C. The purpose is to make IRBs more effective by making sure the right information about investments and progress is available to actually take portfolio decision rather than to simply comply with a process. This point does not detail how much information will be required, but while it is correct that IRBs meeting agendas should be more comprehensive, it is important that the information provided to re-prioritize programs and resources is agile enough to make it usable as a management rather than as a compliance tool.
This point also calls for the use of TechStat sessions, where the health of individual programs is assessed in depth, at the department level rather than just at the OMB level. The intention is to give agencies and departments all they tools they need to improve the effectiveness of their program and portfolio management, without having to play too much of an auditing function from the OMB perspective
22. Redefine role of Agency CIOs and Federal CIO Council (very good)
This point clarifies the boundaries, confirming that Agency CIOs are responsible for managing internal portfolios and the Federal CIO Council is responsible for cross-agency programs. I like the sentence “Agencies will turnaround or terminate at least one-third of poorly performing projects in their portfolio within the next 18 months”, which is a great way to summarize their accountability.
23. Rollout “TechStat” model at bureau-level (good)
This is the natural evolution of bringing TechStat from the OMB level into the fabric of IT management within all agencies.
E. Increase Engagement with Industry
24. Launch “myth-busters” education campaign (good)
When I compared the way the US government has been developing its own cloud strategy, and how the UK government developed its own, I immediately noticed how the latter was heavily influenced by early cooperation with the private sector (in particular with a vendor association called Intellect, which provided free resources from various members to work on the strategy). When I discussed this with GSA and OMB,I was told that it would be very difficult to do something like that in the US.
Well, I hope this particular point of the plan will make things easier, although it is important to keep an eye on how balanced the vendor participation should be. One aspect that is of particular relevance is how the voice of the smaller vendors can be heard, also in order to advance on point 16 above.
25. Launch interactive platform for pre-RFP agency-industry collaboration (very good)
This is another very smart moves, and one that I have been advocating for a long time. Vivek is the right person to drive this, as he did something similar in DC (for the procurement of an evidence warehouse). What is better than having all vendors (and beyond) writing the RFP in the open on your behalf? It is transparent, it makes them aware of your requirements early on, it makes the whole process less expensive and, when the RFP is issued, their response is likely to be more relevant (and more reasonably priced) than it would be otherwise.
My personal take on these 25 points is that they are more or less exactly what is needed to implement the vision of a new role for IT in the federal government. Not all of them will be easy to accomplish within the aggressive timeframe set by the OMB, but even if Vivek takes home only a subset within the current term, he may be able to turn the largest IT spender on Earth toward greater IT efficiency and effectiveness in the long term.
I do hope that other jurisdictions worldwide will take notice and get some inspiration. Actually, this is the first implementation document I have read in a long time that is truly inspirational.