How many units of each?

That's the problem. Sometimes it was one or two, sometimes it was eight or 10, and sometimes it was 15 over a period of time because it took so long to build. You didn't have a production line that was cookie-cutter, where products just go boom, gone, gone. You basically have a build-to-print product that requires a lot of oversight. By the way, that's how most space companies with these types of products operate. The holy grail is figuring out how to productionize the product in such a way that we can keep it hands-off. That's why it's very difficult to do, and that's why those processes are expensive.

What percentage would you say?

In the Navy, 80% of parts were considered sole source. While the Navy could approach other manufacturers for similar parts, once a company was qualified and in production, it was difficult to switch. This is because, in the Navy, you often had to order materials one to three years in advance. Large castings and forgings could take one to two years or longer from order to delivery. This created a significant barrier to entry for new companies. A new company would need to wait for a new platform or a major issue in production or logistics to enter the market. Customers would endure a lot of inconvenience because switching was more painful. However, if a company upset a customer enough, the customer might switch despite the challenges. Balancing pricing and customer management is crucial due to these complexities.

It would be very useful for me to go through an example of the certification process with a specific product within space

In fact, the SLS has already flown a couple of times, and we have some products that still have not completely closed out the qualification requirements, which took maybe six or seven years to complete. The reason for this is that the customer was changing requirements as they discovered new things about the vehicle. Every time they conducted tests or analyses, they would come back with new requirements. We were in the middle of development and had built many parts already, which we had to scrap because of significant requirement changes. The program expanded from the original $30 million estimate, when nobody, including the customer, really knew what they were doing, to over $120 million by the time I left. I don't know what the current figure is, but it could be a $150 million program by now. That gives you an idea of the complexity.