1. 65
  1. 25

    That’s pretty analogous to my experience working as an electrical engineer as well, where orgs that outsource functions to other companies without retaining an in-house expert pay a very high cost, and not just monetarily. They often ship sub-par designs with long delays on top of having high costs. “Buying” can and often does reduce the amount of expertise necessary, but it often doesn’t remove the need for expertise.

    I’m gonna say this matches my experience, too! Story time :-D.

    So about ten years ago or so I was working for a company that developed home automation devices for (what was then the pretty early) IoT market. I was a programmer there, but as one of the two programmers with an EE background, I occasionally got involved in some spark-flying.

    At one point, in the neverending quest for a small, silent, high-efficiency power supply, the company in question contracted another company to develop a reusable PSU module. The existing design relied on a part that became difficult to source, and by that time the company was already focusing on its core expertise which was decidedly not hardware, so the original design team had been let go. We were running on a skeleton hardware crew of one or two people who mostly oversaw production, rather than design activity.

    So anyway, another company was contracted for this, and they came up with what was actually a pretty good design. By that time we had the local skeleton crew in place so I didn’t really know the details, but about an year later, my boss called me to ask me a few questions about, uhh, power supplies.

    It turned out that we’d tried to resell the design to another company (we retained the rights to the design so that was nice) in the hope of recovering some of the money that had been invested. The potential customers, upon hearing how much we wanted for it (which I think was actually less than we’d paid) had a good laugh and said no, and our Technoviceroy wanted my opinion on whether we’d been misled about how complex the design was vs. how much they asked for it.

    So I had a look at the whole thing and it was unsurprisingly a pretty straightforward switching-mode power supply. I asked the other resident programmer who knew their way around a soldering iron (an older and extraordinarily smart hacker that I was 100% in awe of all the time) about it. We both concluded that while the space contraints meant our meager understanding of EMC principles would’ve been likely insufficient to get it right without a lot of trial and error – or quite possibly ever – this was a design we could’ve probably done in-house between the two of us. I think we’d have had to work on it full-time for about four or five months before hitting the original budget contraints, which was likely enough, but who knows. If we’d had an actual design engineer on our team, on the other hand, that would’ve definitely been way more than enough.

    The way I eventually came to think about this was that actual cost of the design was probably about half of what we paid, while the other half was the “yeah, but we can do it” tax. As in, while this is something simple enough that an EE with limited industrial expertise and a programmer with an excellent understanding of electronics could’ve pulled of, you still have to pay for the fact that they’re not actually pulling it off.

    To be clear: not doing it in-house was definitely the right decision quality-wise. While I could’ve probably blundered my way into getting it done, the risk associated with that was definitely prohibitive. Outsourcing it, rather than relying on the limited expertise we had in-house was definitely the correct choice. But the even better choice would’ve been to have in-house expertise.

    1. -2

      Not on topic, but kind of ironic that they have a kernel team, since twitter itself breaks more often than it works, in my experience.

      1. 3

        think of it like a public works program for kernel devs. They got the money to fund it, I’m not going to ask questions :)