Tambrein Bates joined me on the Acquisition Talk podcast to discuss his role in helping build an ecosystem of over 40,000 innovators in support of Special Operations Command (SOCOM). For the last five years, Tambrein has been the director of SOFWERX, and before that spent a career in US Army special operations. During the episode, we touch on:
- The Law of Diffusion of Innovation
- When to use a prize rather than a contract
- The horses vs. jockeys view of innovation
- Moving from a 2-D to a 4-D view of risk
- Whether we set a low bar in defense acquisition
SOFWERX is a non-profit chartered by SOCOM and DEFENSEWERX to transition commercial technology into the military. Tambrein says that commercial R&D has bypassed the Pentagon, and there is a great deal of potential for recombining existing technologies in new and provocative ways.
SOFWERX clients in government tend to access SOFWERX to: (1) help them think clearly about programs/technologies; (2) remove administrative burdens; and (3) access their network of firms and innovators that can address a wide range of requirements from submarines to satellites. SOFWERX is not a “front door” to the acquisition system, but remains important piece to accelerating tech transition.
Podcast annotations
Rapid prototyping has regained favor in the past few years, but some worry whether speed comes at the expense of producibility and maintainability. I asked Tambrein about this:
We talk a lot about design for manufacturing. I think that’s the real difference. When people are prototyping, they don’t consider in the early stages the design for manufacturing component. There’s a big disconnect between engineering and manufacturing across the industrial base, particularly with the advent of the exotically designed tools engineers have at their disposal. They can design just about anything they want, but you can’t manufacture it very well.
Tambrein explained how SOFWERX emphasizes design to manufacture so much that a fabricator was put in charge of their rapid prototyping shop called the Foundry. “That’s how serious we are about making sure something can be manufactured and scaled. That it will be robust and you’re not going to have to redesign it.”
Of course, the manufacturing part is supposed to be emphasized in the traditional acquisition process. Full-Scale Development was changed to Engineering & Manufacturing Design, for example.
In my opinion, it seems like the “stage-gate” approach authorizing development of the 100% solution before the production decision creates a tremendous barrier. Programs only get one shot at a handful of representative quantities. The requirements force an exquisite design. As complications are encountered, quick fixes to stay on cost and schedule create high cost procedures in manufacturing and sustainment.
Many observers, like Charles Perrow in his classic “Normal Accidents,” attribute producibility and maintainability to simplicity. Mary Kaldor argued that Soviet military technology was often more reliable than in the US because it was “uncomplicated, unadorned, unburdened.”
Rapid prototyping and the concept of iterating from a minimally viable product starts you off with a simple design. That doesn’t mean that the product isn’t transformative, but it’s as simple as it can be. As the product encounters the real world, the designers obtain much more rapid feedback on what is producible and reliable. The product then evolves incrementally with a constant reminder of production issues.
By contrast, when production is many years away on the other side of a Milestone C, valuable information feedback is lacking. At any rate, production funding is usually already lined up, and the program will proceed into Low Rate Initial Production. Early LRIPs can often look more like a new round of development than it does early production of a finalized design. Why not just recognize and plan for that?
The concepts of agile development (rapidly iterating on small releases) and devops (where getting development into production and operations is a continuous organizational process) are increasingly relevant for hardware. The difficulties of manufacturing tangible objects, not encountered in software, are starting to be addressed by software. So as more physical processes can be represented and manipulated by software, the more hardware engineering will start to look like software engineering. That’s perhaps a long way off, but I think the trend holds — the shift from tangibles and repetitive work to intangibles and knowledge work.
Tambrein briefly touches on a how the difficulties presented by manufacturing are easing:
There’s been some real advancements on the commercial side, and what’s interesting — I know you’re audience is familiar with 3D printing — that’s another equalizer. I don’t think 3D printing is the panacea, but we can now 3D print things we couldn’t with previous subtractive process.
While rapid prototyping concepts are important, Tambrein reminds us that the traditional linear process still has a role:
There’s three levels of innovation. Incremental, adjacent, and transformational. I don’t think the linear process is always a problem when we’re talking about incremental levels of technology. There are some things that need a linear structured process.
I think that’s right. For incremental advances, there is usually already a large number of users familiar with the product’s features. They can pretty well specify what they want, having already seen many variations. The business processes and engineering are well understood. Tambrein gives the example of boots, but I think cars applies equally well.
However, when the car was transformational, the linear requirements process doesn’t work so well. As Henry Ford famously quipped, “If I had asked people what they wanted, they would have said faster horses.”
For someone who doesn’t have intimate knowledge of the state of the art or state of the possible, they can’t conceive of transformational requirements. Before fission was discovered in 1939 with scientists telling the military what the technology meant, it would have been inconceivable for a staff officer to write a requirement for a bomb with a yield over a kiloton that could be dropped from an aircraft.
We always set the bar a little low I think, meaning we’re not sure what the current state of technology is so we think things are as ready or mature as they are. I tell people, in terms of percentages, about 80 percent of what comes in on the submissions, in general, people have already seen or we have the basic concept so it’s not surprising. But about 20 percent that walks through the door is very interesting, or provocative, or it’s further along than we thought it was.
Listen to the whole podcast for an example of surprising technology, and much more! I’d like to thank Tambrein Bates for joining me on the Acquisition Talk podcast. You can find his LinkedIn post on the Law of Diffusion of Innovation here. Find more about SOFWERX at their website and on Twitter. You can see an interview with Tambrein and Kelly Stratton-Feix on GotUrSix and SOFICTV. Some short articles on SOFWERX are here, here, and here. I’d also like to thank Keri Cline and Christina Caudill for making this happen with their outreach and support.
Leave a Reply