STRANGE REQUIREMENTS: Bradley Fighting Vehicles are good examples of systems with KPPs that ran wild. (Photo by Sgt. Tara Fajardo Arteaga, 113th Mobile Public Affairs Detachment)
The Joint Capabilities Integration and Development System, DOD’s time-consuming requirements development, faces “earth-shattering” change.
by Steve Stark
Later this year, DOD will begin a pilot program that may prove more consequential for speeding acquisition than the revision of the DOD 5000 series of instructions and the implementation of the Adaptive Acquisition Framework.
“This is huge,” said Don Schlomer, one of the authors of a new report from which the pilot grows. “How do I say doing something to the benefit of all DOD? It’s earth-shattering.”
Schlomer is policy manager for Special Operations Forces Acquisition, Technology and Logistics acquisition operations at U.S. Special Operations Command (SOCOM) and earned his doctor of business administration in 2017. There may be no one in DOD who knows more about the Joint Capabilities Integration and Development System (JCIDS). That’s because Schlomer, in addition to working in acquisition and having an active-duty military career, did his doctoral dissertation on JCIDS.
The report that Schlomer co-wrote with four others from the Acquisition Innovation Research Center (AIRC) at Stevens Institute of Technology, “Report on Joint Capabilities Integration and Development System,” was released in September. While DOD set up AIRC, it’s an independent organization.
JCIDS is one-third of “big A” acquisition. The other two parts are the Defense Acquisition System—generally the 5000 series of DOD instructions that describe how to do it—and the planning, programming, budgeting and execution system, which is also undergoing study for improvement. Before JCIDS, the big services developed and validated their own requirements, but then-Secretary of Defense Donald Rumsfeld had begun looking for a more comprehensive way of developing capabilities jointly. JCIDS was born in 2003. The original documentation for JCIDS, Schlomer said, “was 83 pages back in 2003. Now it’s 396 pages. Why?” Not only that, AIRC found that the average system takes 852 days to get through JCIDS, enough time that the technology would be obsolete. That 852 number represents the time it takes to validate an initial capabilities document, then send the ensuing capabilities document through JCIDS and have the Joint Requirements Oversight Council sign off on it.
Today, the vast majority of acquisition programs have to go through the JCIDS process because every program has to be based in an actual need: the requirement. For most, JCIDS is the way they know how to do it.
Congress and DOD agree that JCIDS needs to change, but neither has said exactly how it envisions the capability requirements-development process getting sensible speed, rigor, flexibility and efficiency. That is changing. In the National Defense Authorization Act (NDAA) for Fiscal Year 2019, Congress directed DOD that the secretary of defense would establish extramural (independent) “acquisition innovation and research activities.”
Thus was AIRC born.
In turn, the AIRC report on JCIDS itself was born of the NDAA for the 2021 fiscal year, which directed DOD to “conduct an assessment of the processes for developing and approving capability requirements for the acquisition programs of the [DOD] and each military department,” and figure out how to make those processes faster and more timely.
The Office of the Undersecretary of Defense for Acquisition and Sustainment (USD A&S) is the party responsible, and the one that will establish the pilot project. William A. LaPlante, Ph.D., whose doctorate is in mechanical engineering, is the current USD A&S, and responsible for the AIRC report and its findings.
Now and again scandals over alleged fraud, waste and abuse at DOD streak across the headlines. Sometimes, they really are scandals. Others, maybe not. In 2018, Charles Grassley, a Republican senator from Iowa, wrote an op-ed in the New York Times, “These Toilet Seats Lids Aren’t Gold-Plated, But They Cost $14,000: The Pentagon has to clean up its confusing and wasteful budget,” which decried DOD “paying $14,000 for individual 3-D printed toilet seat lids and purchasing cups for $1,280 each.”
To people who know nothing about how acquisition works, that is scandalous. For those who understand acquisition, the real scandal is key performance parameters (KPPs). They’d ask, “What was the requirement?” Requirements often describe the needed capability in terms of KPPs, which dictate often detailed specifications for the system under development. With Grassley’s $14,000 toilet seat, there may simply have been no other way to stay within the KPPs. That doesn’t mean that Grassley was wrong to be outraged. But he just might have been outraged by the wrong thing.
KPPs, which may or may not have anything to do with interoperability, can force bad decisions, limit tradeoffs and seriously hamper innovation. Schlomer gave the example of the Personal Defense Weapon, which SOCOM was looking to develop. Starting in 2017, it undertook extensive market research. Finally last May, SOCOM found the weapon it wanted. And had tried out four years earlier.
According to Schlomer, if the requirement for the Personal Defense Weapon had a KPP of being 10 pounds 2 ounces, for example, that would stifle innovation. “Well, if it’s 10.4 pounds, but yet it still allows the Soldier to maneuver over obstacles, to be able to carry it in the holster and all that other good stuff, why would you want to discard that weapon just because it’s two ounces more than the KPP? That’s why you don’t want to say [the requirement] that way. You want to say that the weapon needs to be able to be transported, needs to be able to allow the Soldier to maneuver these obstacles, needs to be able to fire at this speed with this accuracy and things like that.”
Conceptualizing prospective systems with KPPs can rule out something radically new. Schlomer said he’d heard that automobile pioneer Henry Ford said, if he’d asked his customers, they’d have said that they wanted a faster horse, not a car. If the best replacement for a system is something wildly different from its predecessor as a car from a horse, it will be hard to predict. Said Schlomer, “You can’t really have a new weapon system that is radically different, that could do everything but was completely new” with KPPs.
EXISTING VALIDATION MECHANISMS
When programs go around JCIDS, there’s rarely if ever a failure of jointness. That’s because the services, in developing systems, understand the importance of systems being joint. They understand the user and by extension the dependencies that the system will have, Schlomer said.
But while the services are looking to middle-tier acquisition more, JCIDS doesn’t work well with it, and that’s a disincentive.
In its analysis of JCIDS, the AIRC team created a model using value-stream mapping to measure the speed of transit through JCIDS. Because documentation of important new capability requirements is classified, the authors could not analyze that directly. They also used the model to look at process alternatives to analyze their speed by comparison to JCIDS.
The report looks at two process alternatives to JCIDS. Both showed significant improvement in speed over JCIDS. Both align with the Adaptive Acquisition Framework’s pathways.
THE MIGHTY SORRD
As it happens, SOCOM and Army Futures Command (AFC) both use simplified and expedited mechanisms that not only eliminate the necessity to use the JCIDS process but are also compliant with the Federal Acquisition Regulation (FAR) and DOD guidance. These are just as legitimate for validating requirements as JCIDS.
Instead of the capability development document (CDD) used by JCIDS, SOCOM uses a mechanism called the “special operations rapid requirement document” (SORRD), pronounced “sword.” AFC developed the “abbreviated capability-development document” (A-CDD) for validating prototype requirements. Programs can swap the 45-page CDD for the 10-page SORRD and A-CDD. In these more streamlined documents, document creators do not lay out key performance parameters. Schlomer said that, unlike the 852-day average for JCIDS, a SORRD approval takes approximately 200 days.
The SORRD, according to the AIRC report, is intended “for expedient approval that aligns with the middle-tier acquisition (one of the six Adaptive Acquisition Framework pathways). The SORRD has a 96-hour limit to be submitted for approval via the Special Operations Command Requirements Evaluation Board, which takes 30 days to validate the requirement.”
In the JCIDS process, the initial capability document (ICD), which precedes the CDD in the overall timeline, “quantifies needed capability requirements and gaps” and “proposes materiel and/or non-materiel approaches to closing or mitigating some or all of those identified capability gaps,” according to Defense Acquisition University.
“The ICD supports the analysis of alternatives and the Milestone A decision of the defense acquisition process. Once approved, the ICD is not updated.” (Emphasis added.) Not only that, it’s not in the same format as the CDD, despite being its basis. The analysis-of-alternatives effort intends to limit wheel-reinvention and make sure that DOD actually needs the capability—i.e., doesn’t actually have something like it already. After that, the CDD—itself a single-purpose document—goes through the JCIDS process.
The coming pilot will use the SORRD mechanism as one-half of a two-pronged approach. The other will use what Schlomer said will be an initial capabilities development document (I-CDD) that will combine the ICD with the CDD but simplify them both. The I-CDD won’t be discarded but updated. Both thrusts will comport with the pathways of the Adaptive Acquisition Framework, and the pilot will be adaptable for use with major defense acquisition programs (ACAT I) as well as mid-tier programs—acquisition categories two and three (ACAT II and below). The pilot will begin with SOCOM, which is an inherently joint environment.
PUT ME IN, COACH
When Don Schlomer read Section 809 of the fiscal year 2021 NDAA and learned that Congress had ordered a report on JCIDS, he wanted to be involved. “I was in touch with the congressional personnel, trying to figure out. ‘Who are you assigning this to, so I can get with them to help them go down that path?’ Because the items that were listed in 809 were the same concerns I listed in my dissertation. So Congress is seeing the same thing I saw when I was doing the research.” Then, he said, “I finally tracked down a commander … from J-8R, who was the gatekeeper, the responsible party.” J-8 is the Joint Chiefs’ force structure, resources and assessment directorate, and J-8R is the requirements part of that. That commander was also the liaison with AIRC at the Stevens Institute of Technology, and it turned out that there was a small group that was developing a model that showed how slow JCIDS was and that, “Yes, in fact, this takes years to get through, and it doesn’t meet the need of the current pathways. It’s just bad,” Schlomer said.
Schlomer was introduced to co-authors Mo Mansouri, a teaching associate professor and program lead for the systems engineering program at Stevens, and Dinesh Verma, the executive director of Stevens’ Systems Engineering Research Center. They were eager to have his participation in the project. “So I became a research scientist under the Stevens Institute of Technology,” Schlomer said. “I got an agreement from the legal here [at SOCOM] that it’s a not conflict of interest.” When the report was finished, Schlomer said, he “also briefed it to the different levels in Congress to explain that, yes, this is important. And I got everybody to agree that we need to, first of all, streamline the process, and then eventually redesign it. And that’s where we are today.”
THIS TIME IS DIFFERENT
In a little more than two years, there have been three noteworthy reports on JCIDS, including the AIRC report. The first was from MITRE Corp. in April 2020, “Modernizing DOD Requirements: Enabling Speed, Agility and Innovation,” which was cited in House conference committee notes on the 2021 NDAA. Second was the one from the Government Accountability Office (GAO), “Weapon System Requirement: Joint Staff Lacks Reliable Data on the Effectiveness of its Revised Joint Approval Process,” from October 2021. That report faulted the Joint Chiefs’ information system, saying, “The Joint Staff cannot assess the JCIDS process because it lacks reliable data and a baseline to measure timeliness. Joint Staff guidance provides a notional length of time of 103 days to review documents in the JCIDS process, but this is not evidence-based.”
All three provide considerable insight into what’s wrong with JCIDS, but the AIRC report is different. Rather than just present its analysis and make recommendations that will never see the light of day, it is intended to effect the real-world solutions it recommends. That’s at least in part because of Schlomer and his experience in the trenches of acquisition, particularly rapid acquisition the way SOCOM does it—not to mention his doctorate and his own military experience (he’s a retired Army lieutenant colonel). Another thing it has going for it is the oversight and participation of the same office (USD A&S) that was responsible for the revision of the DOD 5000 series of instructions and that created the Adaptive Acquisition Framework.
HOW THE PILOT WILL FLY
At the heart of the pilot is the ground truth that “by definition, you have to have a validated requirement,” Schlomer said. “Somebody’s got to agree that, ‘This is what I’m buying.’ Now, there’s no true definition of what that process should be.”
What he means is that there are several different ways to get to a validated requirement, something that vetting has shown the service to need, and that every program has to have. Whatever the process, one that ends in a validated requirement is fine. In the Adaptive Acquisition Framework, the pathways “bypass the JCIDS process,” according to DOD Instruction 5000.80, “Operation of the Middle Tier of Acquisition.” Joint Chiefs’ Instruction 5123.01H also acknowledges that the programs following the middle tier are not subject to JCIDS. That’s in part because mid-tier acquisition, as well as acquisition via other-transaction authority, is intended to end in a rapid, or at least rapider, prototype. In many cases, especially with ACAT II and below, DOD might acquire a prototype capability and then work with the contractor to tailor it to better fit the service’s needs, Schlomer said. That is ideal because with that model, the government is putting current technology into warfighters’ hands, not vaporware that might be possible in the future—or outdated by the time it gets through JCIDS.
The AIRC team used its model to look at alternative processes to analyze their speed by comparison to JCIDS. The report looks at two process alternatives to JCIDS, both of which AIRC modeled using the same framework it used with JCIDS. In a sense, the two process alternatives are prototypes for requirements pathways that work and play well together with the Adaptive Acquisition Framework’s pathways.
Both process alternatives showed significant improvement over JCIDS. Both align with the Adaptive Acquisition Framework’s pathways, and the coming pilot’s two pathways will mirror process alternatives that AIRC featured in its report. “The new processes promise to reduce JCIDS latencies while maintaining the core functions of the process, such as validation of a requirement and consideration for joint interoperability. In general, these new processes eliminate redundant reviews and streamline documentation requirements.” “Latency” is tech jargon for “slowness.” The first process alternative, I-CDD, is for ACAT I programs. The other will be for ACAT II and below, is based on the second AIRC alternative process and uses the SORRD model.
The I-CDD model will merge and shrink the ICD and CDD with an “initial CDD” (I-CDD), the equivalent of the A-CDD that AFC uses. Also at the heart of the pilot is a mindset that one size doesn’t fit all. The AIRC report says, “The current process does not distinguish between the different sizes of the requirements or priority of efforts. For example, the process for approving a new Air Force fighter jet is the same as approving a jungle boot. Senior military commanders might consider that each size requirement document (Acquisition Category, or ACAT, I-IV) should not have the same approval process.” (See related article, “The Jungle Boot” in the April-June 2017 issue of Army AL&T magazine.)
The new pilot documents, SORRD or I-CDD, would exist in a portal, and those who need to comment can do so, but the time has to be limited. “You’ve got some set number of days,” Schlomer said. Reviews would happen concurrently, not consecutively. “If you’re the J-2, you get to see the document the same time the J-3 does, that J-5 does, all that, but you’ve got four days to check it out and make decisions, or check it, [or] make any write ups that you think you need. Four days, that’s it. Absence of concurrence is concurrence.”
There would still be the standard JCIDS process, he said. “If you want to go through a lengthy process to get an aircraft carrier or a huge multi-fighter-type deal? Fine. All right. There’s approximately 81 of them [ACAT I programs] that exist in the country. But for the smaller programs like sensors and radios and all the other stuff like boots, all that?” There is no need for that kind of complexity or high-level approval. “Streamline the approval process down to a one-star or somebody equivalent,” he continued. “Just keep the documents simple. Keep the validation process at an appropriate level and let the services be joint. If they’re developing something, they should be responsible for making it joint.”
The JCIDS process came about because of critical interoperability issues that arose in Operations Enduring and Iraqi Freedom. “DOD established JCIDS in 2003 to provide an integrated process to identify new capabilities from a joint perspective based on the national military strategy,” according to the GAO report. But if JCIDS was intended primarily to assure interoperability from that joint perspective, it has become too much more, accumulating layers upon layers of process barnacles over nearly two decades—requirements for requirements for requirements. That’s not only because accumulating layers of process barnacles is what government programs tend to do, but also because of one of the more fascinating conundrums of DOD: People in the Pentagon who are trained for combat can also be astonishingly risk-averse and seem to prefer getting shot at than having to explain to DOD’s 538-member board of directors (Congress) why a program, in fulfilling a requirement, procured a $14,000 toilet seat or went over budget or didn’t interoperate with other programs. Being a good steward of taxpayer money is stressful.
Key to today’s modernization efforts is speed. JCIDS makes that very difficult to accomplish. The one-size-fits-all JCIDS process is onerous. Speed, however, is also associated with risk. Speeding capability-requirements generation in DOD may increase programmatic risk, but that’s not the only dimension for risk and maybe not even the most important one. The risk inherent in going slow when peer threats blaze along could very well outweigh programmatic risk.
For years, Schlomer has been trying to point out to anyone who would listen that the JCIDS is a very real problem. He is set to retire in September, he said. If, between now and then, he can help DOD halve or quarter the time it takes to get through the requirement-validation process, he will have had an outsized impact on DOD’s modernization efforts.
For more information on the AIRC report, go to https://acqirc.org/.
STEVE STARK is formerly the senior editor of Army AL&T. He retired from service as an Army civilian at the end of February. He holds an M.A. in creative writing from Hollins University and a B.A. in English from George Mason University. As Stephen Stark, he is an award-winning novelist and best-selling ghostwriter.