A version of this post originally appeared in National Defense Magazine in 2012. It has been edited and updated.

In a January 2012 article titled Failure To Communicate, published on the Center For Public Integrity’s iWatch News site, journalist David Axe took a look at the Army’s Joint Tactical Radio System (JTRS).

For those who aren’t familiar with it, JTRS set out to be the Joint Strike Fighter of the radio world, an omni-purpose communications network device that was supposed to do everything but wash your windows.

What it actually delivered – after 15 years and $6B – was little more than a large invoice.

The Army cancelled the JTRS Ground Mobile Radio (GMR) on 13 October 2011, having delivered zero radios. In an explanatory letter from Mr. Frank Kendall, then Undersecretary of Defense for Acquisitions, Technology and Logistics, the reasons for the cancellation included the fact that “the technical challenges… were not well understood due to the immaturity of technology at that time.” Kendall concluded “it is unlikely that products resulting from the JTRS GMR development program will affordably meet Service requirements, and may not meet some requirements at all.”

In other words: Radio development, you’re doin’ it wrong.

Axe’s piece is an excellent story about a rotten situation, but it’s far from the only article about JTRS and its problems. There are dozens of analyses about this ill-fated radio and its aftermath. I mention his article because it contains a painfully poignant phrase that is particularly illuminating.

The phrase comes from Col Dan Hughes, the Army officer who “oversaw ground radio development between 2006 and 2009.” Explaining his office’s goals during those years, Col Hughes says, “We tried to make it [JTRS] better and better and better.”

And that, my friends, was the heartbreaking problem.

To be specific, the problem is rooted in the way project leaders defined better during the 15 years they spent on the project. The word better was treated as a synonym of “more bells and whistles.” The truth is, users would have been better off with… less.

How exactly did the Army go about making the radio “better and better and better”? By increasing its complexity, extending the schedule, spending more money, and making the device larger. They continuously added features and functions and capabilities (on paper), all of which made the design worse and made the users wait.

This wasn’t Col Hughes’ fault per se. The program had been around for a long time before he took over, and during its lifespan there were no doubt thousands of people who contributed to the morass. Col Hughes just happened to be the person who expressed the situation and strategy in perfectly poetic terms.

Much less poetically, seven years before the program was cancelled, a 2005 report by the Government Accountability Office (GAO) observed, with no small degree of disapproval, that the “number of drawings has nearly tripled” in the two years after full-scale development was approved.

That’s right, the blueprint quantity tripled… two years after development got the green light. [Note: The number of JTRS radios in the hands of users did not go up by a factor of three during this time. That number stayed precisely at zero.]

To be clear, the only reason you need a new drawing is if the design has changed. Ideally, a program’s design should be pretty mature by the time full-scale development starts. So that’s a lot of change pretty late in the game. A lot of attempts to make things better which were actually making things worse.

No, not this kind of boxer. 

As the changes piled on, the “Ground Mobile Radio” eventually weighed in at 207 pounds, “several times the weight of existing radios” and solidly in professional boxing’s heavyweight weight class. Recall, the M in GMR is supposed to stand for Mobile, not Massive.

The GAO report tells us the system’s testing was delayed by a year. And then another year. And since that strategy was working so well for everyone, the testing was delayed by a third year. All this on a program that, again according to the GAO, “is proceeding under an accelerated strategy.”

At some point, JTRS leadership realized that constant delays are not exactly consistent with the “accelerated strategy.” So to get back on track, they announced that their strategy “does not allow for testing the radio’s full functionality before initial low-rate production begins.” That’s right, in the name of speed they decided to begin producing the thing before testing it. That’s not fast – it’s just hasty. There’s a difference.

Despite spending $6B over 15 years, the Army failed to produce much in the way of actual radios. Since soldiers obviously still needed to communicate with each other in the interim, the Army ended up buying $11B worth of equipment cleverly named “old-style radios” for use in Iraq and Afghanistan.

I hate to say it, but it gets worse. Brace yourself.

Having already spent $17B, the Army was on track to spend upwards of $12B more to get the radios it needs, according to one estimate quoted in Axe’s 2012 article. I literally can’t bring myself to rigorously research how much the Army actually spent on buying old-style radios or other JTRS follow-on efforts. My admittedly limited research uncovered that in September 2017, “Army officials announced that they were going to put the brakes on the Warfighter Information Network-Tactical communications system, the communications gear that acts as the Army’s battlefield network backbone, after spending $6 billion,” according to an article in Ars Technical

Yikes – and we’re not done yet.

In an interview with Defense Systems on October 4^th, 2011, BG Michael Williamson, then the joint program executive officer for JTRS, explained the results of a Network Integration Exercise held earlier that year. Nine days before GMR was cancelled, BG Williamson put a brave face on the exercise results, saying it “did all the things that we were looking for.”

Then the details came out.

Photo by NeONBRAND on Unsplash

With commendable honesty, BG Williamson acknowledges the radios couldn’t handle the desert heat and had to be shut down. I am sure the Army didn’t want a radio that wilts like a pansy in August, so apparently some things happened which they weren’t looking for.

Sadly, the radio’s problems were not limited to the absence of a huge refrigeration system (which certainly could have been added given a little more time… and money…).

Read a little further into BG Williamson’s interview and we get an excellent explanation of how complexity negatively impacts operations, although we’ll have to connect some of the dots ourselves. Because the radio has so much going on under the hood, it takes a l-o-n-g time for it to boot up. BG Williamson explained:

…even though we had done some work on things like start-up time of the GMR as it runs through all the security checks and initiates the waveforms, we have to go back and figure out if there are ways to bring the radio up much faster. That’s not just because there was a complaint about how long it took, but it’s also because of safety. In the field, there might be a critical message that you need to send.

Indeed, the exercise highlighted the uncomfortable fact that soldiers might sometimes have to send critical messages during a fight. An impatient lot, they may not be terribly excited about waiting ten minutes for the radio to run through a slow series of boot-up processes and waveform initiations. This wasn’t a problem with previous (simpler) radios. Check out how BG Williamson described the soldier’s perspective:

…soldiers came at it [GMR] from the view of the legacy radio, which is, literally, click the button and you start talking, because it doesn’t have the same level of sophistication and the same level of capability.

It’s a striking statement, chock-full of important information. The soldiers expected a simple radio, like what they had before, where they could click a button and start talking. Instead, they received a radio whose “level of sophistication” was out of alignment with their needs. And by “sophisticated” we apparently mean something so fancy it was “ill-suited to the operational environment’s temperatures or the pace of operations for actual users.”

In their well-intentioned quest for simplicity, the JTRS team failed to understand what simplicity is really all about. Yes, having one radio is simpler than having ten, but not if that one radio tries to incorporate every single feature and capability of the previous ten, plus new stuff nobody ever dreamed of. In this case, the result was actually more complicated, not less, and the operational result did not satisfy the users.

And we’re certainly not making much progress or simplifying things if it takes more than 15 years to get that one radio. Delays tend to correlate with increased complexity, and when the delivery schedule continues to slip, it’s hardly ever because we’ve made the thing too darn simple.

The point of this piece is not to dump on JTRS or the Army – that ship has sailed. Instead, when an acquisition program fails – as GMR clearly did – it’s important to take a hard look at what happened and see if there are any lessons to be learned. What wisdom might we glean from the GMR story? I suggest this experience shows that adding time, money, weight and complexity didn’t make the system better. It hardly ever does.

How does this sort of thing happen? It happens when we set out to build a universal system based on contradictory requirements instead of a focused system based on a technically and operationally coherent vision.

It happens when we decide that more is better, rather than insisting on restraint and making the hard decisions about which capabilities are true requirements and which are mere desirements.

It happens when we pay superficial lipservice to simplicity (“It’s only one radio! Simpler!”) while simultaneously indulging an endless appetite for requirements porn (“It’ll handle every legacy mobile ad-hoc networking waveform as well as every future waveform! It’ll have a wide-spectrum amplifier! It’ll be compatible with everything we or our allies use! Simpler!”).

As the Army tried to make JTRS “better and better and better,” they were actually making it worse (and worse… and worse). In the meantime, other manufacturers independently upgraded their existing radios by increments… and actually did make them better. And those better radios actually got fielded and were used.

The result of this approach is inevitably something worse, not better. In the case of JTRS GMR, the final product was far less than the sum of its parts. Don’t take my word for it – ask Mr. Kendall and BG Williamson.

What did the Army need? I’m no radio expert, but it looks like they needed to be able to communicate securely while on the move. Easier said than done, I’m sure, but it’s certainly a reasonable, feasible request. Instead of a capable, reliable, mobile radio system, the Army ended up with a bill for $6B… plus $11B… and a bunch of (new) old radios. Maybe that’s all they needed in the first place. Soldiers would surely have been better off if they had just bought the old radios in the first place, rather than burning through that initial $6B.

But technology marches on, and soldiers still need to communicate, so the Army will have to do something. Thankfully, wiser heads prevailed. Along with cancelling GMR, Mr. Kendall’s letter called for a Non-Developmental Item strategy “with a low cost, reduced size, weight and power variant.”

Then, to make sure everything is absolutely clear, the letter offers this parenthetical translation: “a smaller, more affordable radio.” I wish such a clarification wasn’t necessary, but I’m kinda glad they included it.

The letter goes on to explain that for this smaller, more affordable radio, “minimum required capabilities were reviewed and adjusted. Key waveforms were reduced from seven to two.”

Ah… now we’re talking. Simple, focused, affordable and restrained – that’s a much better set of foundational principles than the JTRS GMR team used in its original trajectory. And that is the real lesson of this story, the application we can put into practice on programs and projects today.

As we face unprecedented challenges brought on by the current pandemic, we’re going to have a lot of opportunities to explore, experiment, and develop new equipment, systems, and solutions. The JTRS experience should serve as a cautionary tale and a reminder that it’s generally a bad idea to launch slow, complex, expensive efforts to develop systems which are not well suited to the operational environments where they are needed. We are far better off using speed, thrift, and simplicity as our guidestars. We are better off doing an iterative series of simple experiments, building prototypes that provide capabilities and provide learning opportunities. Exhibiting speed with discipline, rather than the superficial appearance of speed.

The last thing we need is a repeat of JTRS. The need is too urgent, the consequences of failing to deliver are too serious.  If we understand the JTRS story, it just might help make sure we’ll deliver stuff that is truly… better.

Disclaimer:  The opinions expressed here are those of the authors only and do not represent the positions of the MITRE Corporation or its sponsors.