The Price Of Domestic Just In Time Manufacturing

Hardware is hard, manufacturing only happens in China, accurate pricing is a dark art. Facts which are Known To Be True. And all things which can be hard to conquer as an independent hardware company, especially if you want to subvert the tropes. You may have heard of [Spencer Wright] via his superb mailing list The Prepared, but he has also been selling an unusual FM radio as Centerline Labs for a few years. Two years ago they relaunched their product, and last year the price was bumped up by a third. Why? Well, the answer involves more than just a hand wave about tariffs.

The Public Radio is a single-station FM radio in a mason jar. It’s a seemingly simple single purpose hardware product. No big mechanical assemblies, no complex packaging, not even any tangential accessories to include. In some sense it’s an archetypically atomic hardware product. So what changed? A normal product is manufactured in bulk, tested and packaged, then stored in a warehouse ready to ship. But TPR is factory programmed to a specific radio station, so unless Centerline wanted one SKU for each possible radio station (there are 300) this doesn’t work. The solution was domestic (US) just in time manufacturing. When a customer hits the buy button, a unit is programmed, tested, packed, and shipped.

As with any business, there is a lot more to things than that! The post gives the reader a fascinating look at all the math related to Centerline Labs’ pricing and expenses; in other words, what makes the business tick (or not) including discussion of the pricing tradeoffs between manufacturing different components in Asia. I won’t spoil the logical path that led to the pricing change, go check out the post for more detail on every part. 

We love hearing about the cottage hardware world. Got any stories? Drop them in the comments!

37 thoughts on “The Price Of Domestic Just In Time Manufacturing

    1. Kerry didn’t mean that in the literal sense. It was an exaggerated assertion for fun.

      Is there a word for this type of stylistic device? Brian Benchoff loves to do the same thing.

      And where is Brian? I miss him.

    2. Certainly, by cost per unit. Lots of illusions can be extrapolated from the data.
      You must first solve for the prime industry, manufactured facts. We seem to be edging for top dog there over the last few years.

    3. This is how propaganda works. If you say something often enough you make it true.
      -manufacturing only happens in China
      -you can’t afford to manufacture in the US
      -if you want a successful startup, go to (insert Chinese city name)

  1. I know it’s an art thing that it is only one channel but wouldn’t it be easier to put a tuning knob or switches on the inside so the user can change it but it is not changed as part of normal use?

      1. It has internal switch. But it’s kind of difficult to sell the thing to radio stations as a promotional gadget if that’s something the end customer has to do and is not pre-set in the factory.

          1. (Co-creator of the product here)

            I address this proposal (which we seriously considered) in the blog post! TL;DR it’s *really* hard to find a US fulfillment house that will “tweak” (which in our case means flash firmware, as tuning pots drift over time) electronics and then do final assembly before shipping. FCs want to take boxes off of shelves and put shipping labels on them. The model we ended up with was to get a PCBA shop to do the programming and fulfillment, which works remarkably well.

  2. Earlier this afternoon I was pulling apart my Porlex mini hand coffee grinder with a view to modifying the handle coupling. When I got stuck removing the retaining ring from the shaft, I decided to look around and see if anyone had done the same thing, and came across a website by some guy called Spencer Wright who’d done exactly the mod I had in mind. I still couldn’t figure out how to remove the retaining ring, so in a moment of procrastination came to hackaday, and started reading this article – which was about the same Spencer Wright… Weird coincidence, right?

  3. Many places have 2 public stations or more. Usually a Classical source and talk-jazz as a pair of stations. Our talk 105.9 is weak compared to the classical one, so when the news is on both I switch to the stronger station. Hard to do here.

    On the other hand I have had wearable digital radios that require a tiny slide switch or some such lockout that still fails and they “go off” like a weapon without a safety. Suddenly the news becomes oldies when it “goes off”.

  4. $60 for a mason jar radio? Great concept or not… you can get a lot more radio for that amount. So I guess it’s an art thing. And to be honest, the design concept has its charms. But mason jar radios are nothing new. In fact a long long time ago there was a DIY project for the young radio amateur

    Though in “the good old days” they used a jar for cost effectiveness and simplicity, and not for the “fancy” looks of it. Back in the days a jar was a jar and putting this in a jar was normal. Today when you put something in a jar that isn’t inteded to be in a jar… people tend to call it “design”. Though, what’s design really… name one item made by a human that wasn’t “designed”. “Design” is a word used to sell things for more money. That something is “design” doesn’t mean it’s a good design. And still… I do like the new mason radio though I havent even used, heard or seen it in real life… strange isn’t it? Perhaps there more to “design” then meets the eye.

    1. That link is awesome – I (the co-creator of this product) was not aware of it!

      I should mention that the impetus for using the jar actually *was* for cost effectiveness and simplicity. I’m not sure we ever wrote it up in a blog post, but when we were first starting out we realized that at ~$1 ea (they’re actually closer to $1.30 at low quantities, but whatever) those little mason jars were a lot cheaper than injection molding if you include tool amortization (and shipping, and engineering time, etc). Plus, yes, it seemed like a gimmick that (at least some) people would like. We’ve certainly heard from haters on the design, but overall the mason jar continues to be a very practical choice for our business.

  5. Has this thing some kind of MCU? If yes – there are a plenty of ways progamming it without direct electircal connection and/or switches. E.g. by light (strobe the program) or magnetic (like a card reader). Both of which is cheap and can be done with a minimum of parts. Without MCU – well – even then you could “program” it through flashing some LED. Of course you’ll need a lot more parts. Either way – it is just very unclever to tune the frequency at build-time. Adjust that “detail” in the process and drop the costs by a considerable amount.

      1. I did and the question is still standing. TFA goes: “[…] Josh created a pretty impressive technical demonstration. But finding a fulfillment partner was problematic, and switching to this process would effectively prohibit us from doing the laser marking that our radio station customers love so much. […]” I get that a) you managed to solve the technical problem and b) finding a fulfillment partner WAS problematic. So this problem is solved by now (?). Give them a machine that programs these things. done. So there is only the problem of the jar lid (laser engraving). Easy solution: just use stickers. So no more JIT production. Just produce a large QTY standard-parts and customize it with programming and a sticker. Done. If the engraving is what stands between you and an economic production process – f*** the engraving. It probably not worth it. What are people buying? laser engravings or a radio in a jar?!

        The problem you describe is the whole reason we modularize our products that much. Our products can be combined in several ways, but on the end of the day, a customer needs exactly one configuration. So just-in-time production was once also an option for us. But we modularized and now we produce for stock and just assembly easy plug-and-play modules. It was more expensive during development and during the start, but it monetarized quickly. Our production now usually runs on ~90% capacity.

          1. No. I found your post interesting. All manufacturing stories are something you can learn from (even how not to do it). So keep them going. From an (my) engineering point of view, it is just very unclever to stick to such details like the engraving and having struggles only due to that little detail. But on the other hand: I’m not a sales person; Mabye it makes a huge difference sales-wise if you have an engraving or a sticker. I don’t know. I don’t think so, though. I would get rid of the engraving. ;-)


  6. “Hardware is hard, manufacturing only happens in China, accurate pricing is a dark art. Facts which are Known To Be True.”
    Good try at sarcasm to start off the article, but by reading some of the comments, many people didn’t catch how incorrect those statements are apparently.

    I am a little confused by the “pre-programmed” challenge. The FAQ page on the product’s has simple and clear instructions to set it to any FM frequency, so what is the real drive for having it be pre-set for the end user which makes it a fulfillment nightmare ? I can see the logistics challenge with a laser etched lid for the specific customer’s channel, but the unit itself could be sent unprogrammed. When the customer inserts batteries, they set the channel and bam, happiness.

    The custom etched lid could be done after the fact as well. The etched piece would be a lid that consists of the outer ring area and only a small inner section that goes where the etching is located. Mason jars can easily accept 2-3 lids under the lid ring, so have the laser etched piece sit on top of the lid that houses the antenna, knob, etc. It would be the only custom order item and it can be mailed separately from the unit. All you add is the price of a postcard if you use a folded fiberboard mailer (think cereal box material) and the laser etched portion can be done anywhere that a laser exists with a human that can fold/mail.

  7. 300 FM channels? Where does one go to find that? In Region II with its 200 kHz channel spacing there is only room for 101 channels in the 20 MHz chunk allocated for broadcast FM (though for regulatory purposes they are allocated numbers 200 through 300). Does it fudge it somehow, using Europe’s 100 kHz spacing, plus Japan’s coverage down to 76 MHz?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.