Part 2: Why Good Technology Stalls, and How to Get It Moving
4. When Good Technology Stalls or Fails
Jim Barry
Can you give an example of a technically impressive solution that failed or was significantly delayed? What were the major roadblocks?
Nancy Graham
Steve, you're probably the best to take that one. On the program management side, we have to fix things and make them work. But on the technical side, you've seen things come and go that just didn't get implemented for a variety of reasons.
Stephen Creamer
Most of the technology I've experienced in my 45 years has come in the form of, we can make life better. The initial push after the air traffic controller strike in 1981 was that we could make life better by replacing controllers with technology. The Advanced Automation System was going to address a lot of the needs in that regard. Thankfully, that sentiment changed. It took about seven or eight years, and then engineers started saying we're going to build technology to help the controllers. Then controllers got engaged and started paying close attention to whether this would impact pay or headcount. That friction touches everything you try to do, and there's a real cultural challenge everything has to work through.
The only person I've seen succeed in truly navigating that at a programmatic, national level is Nancy, because with the Oceanic system she literally changed the way controllers worked. It took around seven years. Controllers received eight weeks of training on how to use the computer, what it did with the information, and how they interacted with it, so they could trust that the computer's decisions were being made correctly. That system has been in use for the last 20 years. That's the example of one that went really well, and it succeeded because representatives from the facilities were engaged with it and invested in seeing it succeed.
A program that hasn't succeeded has usually failed because that commitment wasn't fostered to the level necessary to gain critical mass. The Advanced Automation System collapsed under its own weight for that very reason. Data Communications was a great technological leap that only recently got into air traffic control centers across the United States, and only with suitably equipped airplanes, which aren't the majority of the fleet. We're still working on getting it adopted to the point where it can be utilized by everyone.
Jim Barry
Why did it take so long?
Stephen Creamer
In 1990, when it was created, the technology probably wasn't ready. We were using very low-bandwidth connections to the airplane. Today you have Starlink, although we don't use Starlink yet for air traffic control, and that's a regulatory limitation I think we're about to overcome. The critical mass is almost there.
As connectivity has improved, things that have become routine in other sectors haven't been adopted into the aviation system, because the cost of adopting change is incredibly high. That's why the US air traffic control system atrophied for roughly 20 years. There wasn't a proactive approach to identifying new technologies that could be implemented, because there was so much cost pressure that people gave up looking. The FAA's approach to program delivery and prioritization became a form of triage in the latter NextGen years. That became self-defeating.
It takes me back to what I said at the beginning: what are you trying to achieve? Right after I first got to ICAO behind Nancy, we'd been talking for 10 years about how to implement change promoted by the workforce, then creating a method to get that change picked up and moved through the system quickly enough to deliver it and keep the momentum going, at a lower cost because people commit to it.
These were Nancy's ASBUs, the Aviation System Block Upgrades. What's the problem? What's your idea to solve it? Let's talk about that idea with all the stakeholders. Do they all think it's a good idea? Then let's test it and see if it might work. If it still looks promising, bring the regulators in and ask whether they would approve it in this form. After that, what's the cost-benefit? If you've done all those steps, you've already had to refine it somewhat to get through them. Does it still have the same cost-benefit you thought you had at the beginning? If it does, let's move forward. If it doesn't, put it in the learning pile.
That approach borrows a lot from what startups themselves have taught us. But in government, nobody makes a program to fail, that's almost something we don't accept as part of the culture. I think we're getting there now, though. I see the current administration, because of the way the new air traffic control system procurements are set up, is actually willing to push some things and try them, and that needs to be fostered.
John Illson
I've certainly seen a lot of the same things Nancy and Steve talked about. The one thing I'd add is that new entrant companies are so focused on the technology, as they should be, that's how they develop such impressive solutions. But they look at the collaborative process and the regulatory process as afterthoughts. They know they'll get to it eventually, it's one of those housekeeping items to them.
If they're not genuinely invested in that part of the business, they don't give it enough time. When they're told no, as Nancy said they will be, it's a real setback, and they're surprised. Sometimes their investors are surprised. They might lose valuable people who want to move fast, and things start to unravel. I've seen a lot of programs significantly delayed because that engagement was never a line item in the project plan. They didn't work backwards to figure out how far in advance they needed to start engaging with the regulator, government stakeholders, or industry partners to get across the finish line. It takes so much longer than anyone expects, because the technology often moves faster than the process.
5. Collaborating with the Regulator Is the New Entrant's Job
Nancy Graham
I want to talk about some positive things, because so far we've talked about what's in the way. There are some really important things happening now on the positive side. In the past, all the money came from the government, the idea of private investment just wasn't part of the frame. Same for space, it's largely private investment now.
In the early 2000s, it appeared to me that private capital might be a one-time infusion. I don't think that anymore. I think there will be a continued external cash infusion from outside government, and that's a good thing. Because that money comes from outside, it's more important that we learn to talk to investors as well. These days I spend time with investors trying to explain, this approach is sufficient, or this one isn't quite there and here's why, but here's what we can do instead. How do you work with investors so they're not caught off guard when things go wrong? They need to understand how the system actually works.
There aren't many people doing that, and maybe part of what your company can do is help investors understand the processes it takes to obtain certification and to be patient as companies work collaboratively with the regulators to go through open questions and deal with the necessary setbacks that allow for learning. The FAA is extraordinarily capable, but it has significant staff turnover. It's many times worse around the rest of the world. You have to factor that in, and these companies are all international companies now.
When things go wrong, you have to back up and take stock, because that's not how we traditionally learn in this industry, but it is how Silicon Valley learns: fail fast, but in a contained way. The aviation side has learned to use laboratories, sandboxes, and trial programs to do this in a low-risk environment that gives you information to begin assessing risk. My coaching to technology industry newcomers is, you have to let the regulators learn with you. You can't just tell them what happened. These are visual people, pilots and air traffic controllers, they want to see this work. They're not interested in your documentation alone. They need to see that it works, because for them it's deeply embedded in their ability to do their job, and their ability to know that they're not putting anyone at risk.
The industry has to accept that the education process is its own responsibility. Again, is it the name of the project? No. But it's a fundamental part of the job. Regulators hold the public trust, and they are resource constrained, rightly focused on passenger transport and established operations. They often haven't had direct exposure to the new technology yet, or the industry is proposing something that simply isn't regulated today. You have to work through that together. At the working level, the incentive structure tends to favor caution, which makes sense given what's at stake. So my suggestion to new entrants is, don't look for a yes, you're unlikely to get one early. Look for a not-no. Then you can work your way forward from there.
Jim Barry
It sounds like you've given that advice to a lot of companies over the last few years.
Nancy Graham
Yes, at least 10.
John Illson
I've certainly seen a lot of the same things Nancy and Steve talked about. The one thing I'd add is that the collaborative and regulatory process is something companies need to take seriously from the very beginning, not treat as a later-stage item. When they're told no, as Nancy said they will be, it's a real setback to them and to their investors. Those that have built the engagement early, who've treated it as a core part of their development process, are the ones who come through the other side.
