Issue link: https://iconnect007.uberflip.com/i/1265351
26 SMT007 MAGAZINE I JULY 2020 Matties: As long as profit margins are so thin, if you have the right partner, there is no reason to seek out a captive facility. Kelly: It's also the reason you don't want to move too much because the transfer and move- ment costs can kill 3% pretty quickly. That's why you want to get a good partner and keep them. Matties: Maybe it's quicker for a captive facil- ity to be optimal. What trends in technology do you see in the market that the industries should be mindful of right now? Kelly: The number one trend is digital transfor- mation. I know that sounds like a buzzword, but if you look at the level of sophistication, it's quite low. There's still a lot more to be done. Again, we have to partition this. Silicon semiconduc- tor wafer technology is probably somewhere between 10 and 15 years ahead, so I'm not talk- ing about that; I'm talking about EMS in electron- ics manufacturing, which includes PCBA, PCB, mechanical assembly, system fab, and final test. The digitization of operations and chang- ing the way people work alongside that digi- tal transformation is huge. I know that sounds vague; everything is about digital, but there's a lot to it. It's how designs are made at the very beginning. Someone talked about design for manufacturing. That's still not done very well. Some people have figured it out, but at large, it's not designed for sustainability or some of those back-end processes. The digital aspect of it enables business-to-business, so transferring data securely is number two. As soon as we talk digitization, it's probably in parallel. Security is right there. In the elec- tronics manufacturing industry, if we're going to make real strides in implementation—not just describing these great technologies but making them work—we need common digital platforms. An example is IPC's Connected Fac- tory Exchange (CFX) protocol, where we have a common language that allows data to be moved around. If we can't agree on the com- mon digital thread, then we can't connect the way that we want to. were in more of that conventional CM model. The OEM owned the design, they partnered with someone to build it, and you know the rest of that chain. Over the last five to eight years, there has been more and more expan- sion happening by the EMS providers. That's how ODM was born, where they basically say, "We'll finish off the design for you. The OEM needs to own the design, but we'll fin- ish up physical design, or we'll do the roll- up-your-sleeves-type work on the back end of design." The ODM is trying to capture more of that work upfront before they build anything. They're trying to open up the aperture to the left, closer to design. They're also offering more final system build services at the back end, so instead of produc- ing a PCBA, I'll make it up—10 PCBs come to a hub and they get put into a subsystem, bolted together and made into the final assem- bly. The EMS is also doing that now. I'm not sure if you've heard the terminology of a level build, like L6 versus L9, where that EMS pro- vider is producing the final subsystem, includ- ing software and firmware loaded. They're try- ing to expand to the right as well. When an EMS now says, "I'll be your partner," they're expanding their service offering to the left on the design side and to the right on the final ful- fillment side. By that, they're trying to expand their margins. Matties: That makes sense. They have to keep adding value. Kelly: Remember that with EMS, margins are razor-thin—around 2–3%. If you hit 10%, it's a good financial result. Over the last five to eight years, there has been more and more expansion happening by the EMS providers.