The time is right for a public policy discussion contrasting the SWOT for fully outsourcing government IT - no longer would government own, maintain, and upgrade IT infrastructure or application development, but instead private industry would provide government with secure cloud-based DaaS/SaaS that would accelerate commercial as well as government security, convenience, affordability, and trust compared to ownership. Identification of intended and unintended consequences of a change of this magnitude and strategic actions, technological capabilities, and timelines necessary to transition from "owning/running" to "using" would be part of the discussion, as would exceptions to outsourcing. My premise is that government involvement in IT technology will eventually follow the model for weapons technology by the military:
I posit that for weaponry, the Department of Defense has segmented expertise between private industry and government into these broad roles: the government presents operational needs as scenarios or statements of objectives (SOO) and specializes in the validation and acceptance testing phases, with few of its resources dedicated to weapons design or development. Private industry specializes in responding to government requests (expressed as contracting opportunities) as well as developing and pitching unsolicited inventions (SR-71, BAE Systems Bofors 57 mm 3P with all-target programmable ammunition, etc.). This specialization was the unplanned result caused by the complexity of weapon technology – the government expertise was in fielding and logistics but not in knowledge of emerging technological capabilities and what those technologies could achieve when applied to weaponry.
Conversely, I posit that we are in the transitions stage from government expertise to private industry expertise with respect to IT technology as it realizes the functions of government, with a corresponding need to focus government resources on specification of scenarios and objectives coupled with validation and acceptance testing, while private industry assumes the role of inventor and developer. This realization was driven by the publication of a list of government IT projects that struggled, not the least of which was the implementation of the federal healthcare exchange – perhaps where for the first time a single government website would experience the same high volumes, SLA expectations, and scrutiny as various commercial websites (Facebook, Twitter, Amazon, Google) and government personnel were not the experts in construction of a massive system of such capability, especially when constrained by requirements to operate with multiple legacy systems.
If the hypothesis is correct (government and private industry have specialized in roles and resources for at least one type of complex system where neither side is capable of understanding the opportunities or activities necessary to maximize realization of utility – I cite weapons development and deployment) and the model is valid (that the same forces that forced the specialization in the first case are beginning to occur in the case of systems of information technology when it realizes significant government “product” – then consequences for the future are to plan for the inevitable reality that government-run IT is not comparable to private industry IT; that the future for massive systems of citizen interaction will only be possible from private industry and government’s role will not be to specify, buy, maintain, and refresh the equipment but to determine the use models, not to construct and operate the data centers but to validate the operational requirements and SLA and to ensure the deployment to the target audiences through oversight. The implications are to move away from setting up, sustaining, operating, refreshing, and supporting IT systems to technical expertise in overseeing implementations – an outsourcing, as it were, impacting what agencies who currently operate citizen services in data centers should be doing to exit from the day-to-day operations and become change agents for new citizen services and acceptance testing for systems that provide those services.
I also believe that IT procurement reform (modeled after defense acquisition reform) is insufficient; better buying is not the problem; requirements and architectural choices that underlay procurements are the critical success factors lacking in government IT efforts. Summarizing:
1. There exists in government a model (weapons acquisition) of specializing expertise into design/develop/produce (industry) and into use/doctrine/validation/acceptance (military)
2. That the existing weapons procurement model is reasonably effective, although not perfect as-is (hence repeated procurement reform legislation/initiatives)
3. That IT systems/solutions are/will become as critical to realizing agency mission objectives as weapons are to realizing military commitments.
4. That IT systems/solutions as presently executed by government are passable–but the trend is that IT is a critical mission tool, but IT is not a critical expertise priority and costs will spiral upward.
5. That a “weapons model” could be applied to “information technology systems/solutions” in an improved way so as to minimize the cycle of procurement reforms accompanying that model.
6. That government cannot attract and retain the type of technical expertise required to excel in design of state-of-the-art IT systems and applications. Only 4 percent of people who hold degrees in computer and mathematical sciences were employed by the government as of 2008, a Ford Foundation report prepared by Freedman Consulting says, compared to 70 percent in the private sector (http://www.fordfoundation.org/pdfs/news/afutureoffailure.pdf).
Instead of better acquisition regulations:
1. Government would facilitate a public debate regarding the strategy of outsourcing IT coupled with an enhanced procurement practice modeled after the specialization of duties and responsibilities performed by the military and private industry in weapons acquisition.
2. Policy-makers would be knowledgeable about the implications for mission-critical IT technology development, proper roles for government and industry, and active management of the consequences within government. Going forward, technology development and operations would be managed and improved as a strategic procured product/service.
3. An outcome would be analysis contrasting “as-is” IT development and procurement practices against outsourcing all IT equipment and development, illuminating likely costs and schedules to achieve the scale, security, reliability, and performance required from industry, the critical strategic activities necessary to implement outsourced IT practices, and the implications outsourcing IT on the existing government IT infrastructure, duties, and workforce.