Failure and Success in Defense Procurement. Also, Revolutionary Versus Evolutionary Development

The DoD has shown a remarkable inability to effectively manage major weapon system procurement programs over the last 20 years. Browsing the web today, I find two articles that highlight just how bad the process is.

The first is an article in DefenseTech displaying some of the most expensive failed programs of the last decade:

Future Combat Systems (FCS) $18.1B

Comanche helicopter $7.9B

nPOESS satellite $5.8B

VH-71 Presidential Helicopter $3.7B

Expeditionary Fighting Vehicle (EFV) $3.3B

Transformational SATCOM (TSAT) $3.2B

Crusader $2.2B

Advanced SEAL Delivery System (ASDS) $0.6 B

Armed Reconnaissance Helicopter $0.5 B

Aerial Common Sensor $0.4 B

CG(X) next Generation Cruiser $0.2B

CSAR-X $0.2B

Most of these failed programs fall into two main categories. First, there are those programs that were simply too technologically aggressive to be produced in any economical fashion, in anywhere near the numbers of platforms required. Other programs such as the RAH-66 Comanche and the Crusader howitzer programs were simply overtaken by the pace of technology. That is, they were designed for a networked environment before the rest of the world even started to think of that level of connectivity, and the development programs were so long that by the time they were ready for production, the state of the art had actually long passed them by.

The Dew Line, by way of contrast, has a list of successful programs (in this case, limited to aircraft procurement).

1. UH-72A
2. EA-18G
3. CH-47F/G
4. MC-12W
5. H-1 Upgrades*
6. MQ-9 Predator

The first thing that leaps out when you look at the list of successful programs is that there isn’t a single “transformational” or “revolutionary” platform on the list. Each program has built either on an existing airframe, or is an evolutionary adaptation of an existing technology. Program managers were careful to not add every possible feature to the design, and locked in the basic design early on. They worked to ensure that the requirements were realistic, and that the state of the art could reasonably be expected to fulfill them in a timely and cost effective manner.

I’m greatly concerned that the DoD can’t seem to take the same approach to other major weapon system procurement programs. Problem programs such as LCS, FCS, the F-35 and EFV all seem to fail whenever “spiral development”shows up,  where basic research and development of underlying technology and the weapons design and procurement process are intertwined.

I’m not arguing that the military shouldn’t pursue new technologies or only buy revamped versions of existing designs. But it is apparent that unless critical design technologies are relatively mature, the DoD cannot be expected to incorporate them without massive cost overruns. And who thinks that’s a viable option in the face of austere defense budgets looming ahead?

*I’m not at all sure I’d personally include the H-1 Upgrades as a successful program (other than to admit that aircraft are being bought).

7 thoughts on “Failure and Success in Defense Procurement. Also, Revolutionary Versus Evolutionary Development”

  1. I am thinking an unfinished thought that the navalized F111B was a 1960s example of “spiral technology” in which the AWG-9 radar/Phoenix missile combination was not ready for prime-time but since the aircraft got killed off, the system was ready by the time the F14 showed up to utilize it.

    1. Well, the AWG-9/Phoenix was actually a development of the AIM-47 for the YF-12, so it wasn’t all that spiral (plus a lot of development that went into the Eagle/F6D Missileer program).

      So while it was a formidable development program, it wasn’t revolutionary.

  2. NASA procurement is much the same. Ares was supposed to be evolutionary but ended up with more R&D than the budget could bear.

    I also wonder about the product data management systems and whether the choice of a poor PDM dooms a program no matter what. What made FCS fail?

  3. At least as far as aircraft development is concerned the price keeps getting driven up by not freezing the design and the customer continually making changes. The Army was it’s own worst enemy in the Comanche debacle.

    The folks in acquisitions needs to require a “plug and play” capability because in today’s world the pace of electronic/avionics development is much faster than the pace of new airframe development. A Block 3 Longbow is nothing like an A Model or even Block I, but because of Boeing proprietary issues it takes forever to make changes in electronic/sight upgrades etc.

  4. Outlaw13 nails a key reason for these issues…requirements creep on the part of the requirements definition community (in the Army it is TRADOC). If the combat developer lets the good idea fairies to thrive you will key a program because there is no way for the acquisition community to get a product built on schedule and within budget.

    Part of the problem is weapons to fight the last war (Crusader, EFV, CGX), part of the problem is there is a lot of the pushing of the envelope (FCS, TSAT) and part of the problem is guilding the lilly (pretty much all fo the rest of them.

    I will say there are some successful developments entering the force thanks to soem of these programs. The Army is getting some useful items subsystems and components out of the FCS which is really helping the Stryker program.

    As for some of the proprietary issues…during development you try to get the contractor as opposed to the government to accept most of the risk. So you make the contract a cost plus. When you go to production you normally transition to a fixed price with incentive fee. This helps keep some of the proprietary costs down but not all. In complex systems it is still going to occur. ..and why 70% of a systems life cycle costs occur after production is complete and not during development.

    And don’t get me started on operational test!

  5. It is rare that something comes along that is revolutionary. The airplane, then the Helo were such things. In Surveying, the Total Station was evolutionary (theodolite combined with electronic distance measurement that came from a combination of the transit and the result of speed of light measurements in 1947), but GPS was revolutionary.

    Revolutionary things cause a sea change in the way things get done. Evolutionary change results in things being done more efficiently. Airplanes allowed increasingly large explosive packages to be delivered over much longer distances than artillery could deliver them. GPS did away with triangulation surveys and all Geodetic Surveys are now done by GPS. A single trooper can carry a hand held GPS with him and get his own position to an accuracy that would have required days of work before.

    Methinks the military has gotten caught with the revolutionary instead looking closer to home. that doesn’t mean Comanche and Crusader are failures even if they did not reach operational status. The example of the Phoenix missile above is a good one. The lessons learned can lead to completion down the road, or can be incorporated into later successful systems.

    We can predict fairly well the direction of evolution in systems design, and there fore, military hardware. Evolution is almost impossible to predict. Science Fiction writers imagined certain things coming about, but the effects of those things usually weren’t even conceived. I saw the old Transit system (the Navy’s early Navigation Satellite system). I could see the seeds of GPS in it, but did not imagine a Survey grade system that could Survey the entire US to an accuracy undreamed of by early geodetic Surveyors, in a time period so short it would seem miraculous by those early surveyors. Myself and several hundred others gathered the data needed to redo the geodetic resurvey of the country in less than 5 years. The Coast Survey, or the later Coast and geodetic Survey couldn’t do the Southeast in 5 years.

Comments are closed.