The Pentagon rebuilt acquisition for speed. It still needs a way to discover, validate, own the problems worth solving
Editor’s Note: A version of this story previously ran on Real Clear Defense.
Ten years ago, Steve Blank and I made a bet. Put a real problem, owned by a real person in uniform, in front of a smart team. Force that team out of the building to talk to a hundred people who live with the problem. You will get a better solution, faster, than the requirements system produces. That bet became Hacking for Defense.
This summer I watched the strongest teams to come out of that pipeline go through a ten-week commercialization fellowship run by the Defense Innovation Unit. They got advisors, a modest prototyping fund, and access. A handful came out the other end with a named government customer and a real path to the field. One holds an Army authorization to run on soldiers’ devices in two forward-deployed units. One is producing intelligence a combatant command has entered into the official record of its annual exercise. One signed a research agreement with an Army lab in week two and had Ukraine’s counter-drone accelerator ask to set up a live shoot-down test.
The rest had comparable talent and, in several cases, better technology. What separated the few from the many was not the product. It was the problem.
I have spent thirty years watching the Pentagon try to buy its way out of that lesson. As director of the Army’s Rapid Equipping Force at the height of the counter-IED fight, I learned that a warfighter’s need cannot be precomputed from a headquarters in Virginia. What I saw this summer is the cleanest evidence I have had since that the same rule governs every startup, lab, and program office that wants to put capability in a soldier’s hands before the enemy adapts.
Looking back across the cohort, three things separated the teams that reached a customer from the teams that did not, and all three were visible before a dollar of prototyping money moved. All three are about how the problem was recruited.
The sponsor suffered the problem rather than observed it. The sponsors behind the teams that made it were a lieutenant in a cavalry regiment, a captain in an engineer brigade, and a company commander. Nearly every problem that reached the field came from an operational unit. The problems that went nowhere came from R&D centers and portfolio offices, from people who could describe the problem but did not carry it, and who proved far less inclined to kick down doors afterward. The lieutenant became the champion. The portfolio manager became a contact.
The stakes were measured in lives, not efficiency. The strongest problem statements read “under direct enemy fire” and “shut down at will by the Chinese Communist Party.” The weakest read “slowed design cycles” and “manual workflows.” The language turned out to be a leading indicator of which sponsors would champion a solution. I learned that one the hard way.
While I was running the REF, I was called into a Congressional office to explain why I was spending money on “green energy” in Afghanistan. I told them I was not trying to bring down the price of a gallon of gas downrange. I was trying to cut the number of resupply convoys that had to move that gas over roads seeded with IEDs, because every convoy we did not run was a truckload of soldiers who did not get blown up. A month later the defense authorization bill carried a new $20 million line for REF’s power work in Afghanistan. Same technology. Same program. The only thing that changed was that the problem was finally stated in the currency that matters. If the problem statement does not name who gets hurt when it goes unsolved, nobody will fight for the solution when the calendar gets hard.
The team validated its hypotheses in conversation, not in the lab. One team arrived with an app for navigation. Sixty interviews with soldiers told them planning, not navigation, was the bottleneck; that connectivity could never be assumed; and that units reject standalone applications on sight. They threw the app away, rebuilt it as a plugin for the tactical software soldiers already run, and added a concealment mode no fielded NATO or commercial system offers. That team cleared an Army Risk Acceptance Memo weeks later, and its plugin was authorized for soldiers’ devices. Another team spent five weeks unable to reach its own sponsor. Every team that stalled was short of the same thing: a dated test with a named customer waiting on the other end of it.
Steve has been saying “get out of the building” to founders for twenty years. This is what it looks like when a founder does it with a platoon.
The Department of War has just executed the most ambitious acquisition reform in six decades. JCIDS is dead. Portfolio executives own requirements, contracting, testing, and sustainment end to end. The new system rewards speed to delivery. I have argued that the department built a faster engine without building the steering, and that three of the six phases of any working innovation cycle, detecting the problem, assessing whether the solution worked, and distributing what one unit learned to every other unit, still have no owner.
That fellowship is what steering looks like at the smallest possible scale. A problem pulled from an operational unit. A sponsor who bleeds from it. A team that tests its assumptions against the people who will use the thing. A named champion who opens the next door. A solution that drops into what the force already runs, with no policy exception and no waiver. Ten weeks, and a lieutenant’s problem became an authorized capability in his soldiers’ hands.
Now look at where the department’s own requirements still originate: headquarters staffs, service-level assessments, and portfolio offices reading studies – the same sources that produced the problems that went nowhere this summer. We have spent the last year making the middle of the pipeline faster and left the front end exactly where it was. Faster procurement of the wrong requirement still loses the next war.
The Rapid Equipping Force and the Asymmetric Warfare Group had teams forward with operational units whose only job was to find the problem before headquarters wrote the requirement. Those teams no longer exist. A few handfuls of university students with plane tickets to forward-deployed units are now doing that job.
The fix is not another authority or another fund. It is discipline about where problems come from, applied everywhere the department and its industry partners decide what to build.
Source problems from the people who carry them. A portfolio executive at headquarters, however empowered, cannot see the distinctions that matter without ground truth from the fight. Stand up small, cross-functional problem discovery teams with operational units, scanning every week for how the adversary adapted since the last one. The department once had them. It should have them again.
Make a named champion a precondition, not a hoped-for outcome. Before money moves, a problem should carry the name of someone in the unit who will make the introductions, sit through the field test, and sign the memo. If no one will put their name on it, it is not yet a problem worth solving.
Require a dated test with a named customer before a prototype is funded. Not a demo day. Not a pitch. A date, a place, a unit, and a pass-fail condition. Every team I watched that had one converted its money into evidence. Every team that did not, did not.
Build for what the force already runs. The fastest teams wrote plugins, not devices, and signed research agreements instead of proposing programs of record. Nothing in their path needed a waiver. Industry should hear that clearly. The path to the warfighter is through the software and the agreements already in his hands, not around them.
Einstein said that given an hour to save the world, he would spend fifty-five minutes on the problem and five on the solution. For a decade the department has spent its reform energy optimizing the five minutes. That lieutenant did not wait for a requirement. He handed a team his problem, and ten weeks later his soldiers had the solution on their devices. Getting technology to the warfighter when he needs it starts there. It always has.
