As Federal chief information officers (CIOs) continue to face challenges in modernizing and securing our nations’ technology infrastructure, many are faced with a with a decaying technology foundation that is often decades old, and customized.
Add to the mix a set of bureaucratic rules dictating how they buy IT, and it’s nearly impossible for them to upgrade to the modern, best-in-class technology they need.
A recent report from the Alliance for Digital Innovation describes these issues in sobering detail:
…“In recent years, how much of the government’s IT spend has been wasted? Studies show the average enterprise-spend for IT, across all commercial industry sectors, is about 3.3 percent of revenues. If the government IT spending matched the industry average as a percentage of revenues, the government would have saved about $717B since 1994. If the government had been able to keep its spending at [Senator William] Cohen’s rate of 5 percent (the same rate already bemoaned at the time as wasteful), it would have saved about $345B in the twenty-five years since Computer Chaos was written, and FASA was enacted.”…
The amount spent on maintaining legacy systems versus modernization is staggering:
…“Despite these profound losses to American taxpayers, IT funding continues to flow from Congress at a rapid pace—much of it destined for the operations and maintenance of antiquated legacy systems or slated for custom IT development. Deltek’s recent “Addressable Government IT Market” survey for the six-year period, between 2018-2023, forecasts $664B in net new IT dollars to be spent government-wide. To be clear, this is 66.4 percent of a trillion dollars to be spent over six years—almost two times the market valuation of Exxon Mobil—and only a couple of years of spending less than the money required to buy Google or Microsoft, in their entirety. Of this enormous six-year figure, only $78B, a mere 11 percent, is predicted to be spent on new commercial software. Meanwhile, $191B, the lion’s share, is destined to go to an entrenched system integrator community, to be spent on IT professional services, where most of it will fund the maintenance of a patchwork of ancient bespoke non-commercial systems.”…
One of the significant barriers facing federal technology is the federal procurement process itself.
IT procurement rules are changing, and several principles need to be a focus to help increase flexibility and allow for rapid, agile technology procurement to modernize the federal IT landscape.
First, government technology should not focus on custom-built solutions. The “we are unique” excuse is weak, as the federal government needs to buy be best-in-class technology, most notably existing, off-the-shelf commercial platforms. These solutions have been proven at private-sector firms of similar scale and complexity to government agencies. A focus on digital services procurement is critical to changing this mindset, especially eliminating the massive, billion-dollar, waterfall-based IT programs that are destined to fail. Further, these programs normally come with massive costs via thousands of hours of added “consulting” or “services” fees. What are the fees for, anyway?
Automation is central to these solutions, from data analytics to security. Today’s best-in-class technology (e.g. distributed ledger, artificial intelligence, etc.) can automate most of those tasks in seconds, eliminating redundancy, greatly reducing error rates, and lowering instances of waste, fraud, and abuse. The world’s largest banks, which are not all that different in size from certain civilian agencies, and face similar issues, are using modernized platforms that give them accurate and complete data across their enterprise in seconds.
Certainly the Office of Management and Budget push to have agencies use best-in-class contracting mechanisms is helpful, as is the scaling of the Digital IT Acquisition Professional Training (DITAP) program. However, user-centered design, agile software development, and paying for outcomes and results should also be part of the mix. Hiring firms to write 100-page requirements documents for massive technology buys is a recipe for failure, as is creating custom blanket purchase agreements that take years to create.
Procurement must become more agile and customer-focused than it is today. It ought to take weeks or months, not years, for agencies to get into production the technology being purchased. One of the significant obstacles to technology implementation is the challenging authority to operate process at most agencies. By the time technology is typically deployed, new threats have emerged, and original requirements are obsolete.
Technology contracts must also be flexible and eliminate locking in the status quo technologies for years on end through standard “base year plus four (4) options” mechanisms. Prototyping and demonstrations of technical efficacy need to be used more often, such that to tools can be evaluated for a shorter period of time, and either further developed, or purchased as a subscription service on a consumption-based model. These contracts are for shorter durations, with specific factors for further investment or cancellation. This model lowers risk and prevents the typical situation of being stuck with technology that does work, and locked into the contract requiring waste, or requiring difficult processes to terminate.
Parts of government are moving in the right direction. Models such as the United States Digital Service and 18F demonstrate the value of more commercial-like development and implementation of technology, and how running a federal program more like a business makes sense for the taxpayers.
It’s time to continue this momentum and normalize these successes across the government. We simply can no longer afford to waste billions just keeping the lights on. We must modernize our federal IT into the 21st century, and adapt to the changing technological landscape that is leaving us behind.