2021, Chapter 15:
Weirdest. Year. Ever.
I’ll admit it: I thought last year was weird. But 2021 was definitely the weirdest year ever.
Here’s the thing: in 2020, we had quite a bit of already-built stock, so the main challenge was just shipping things. And keeping people healthy. And scaling up to meet demand, as people working from home pushed sales higher. And wondering if California would be shutting us down any time soon, hence the search for a second facility in another state.
Ah heck, 2020 was crazy too.
But I didn’t expect 2021 to get even weirder. At worst, I figured, we’d have to deal with unexpected stuff from the ESS transition as we got back to normal. We’d be able to launch a bunch of cool new stuff and it’d be a super-happy year.
Instead…we got boned 4X more than expected in 2021.
So am I complaining? No. Despite the, ah, unexpected “learning opportunities,” AKA “getting screwed,” despite the fact that we weren’t sitting on large amounts of stock like in 2020, despite starting up an entirely new manufacturing facility in another state, despite a relative dearth of new products, we managed to grow. Pretty significantly, in fact.
But am I tired? Yes. Not tired like “heck with this, sell the company!” tired, but tired like, “Hey, can we quit with some of the parts shortages so I can do the fun stuff of designing new products,” and tired like, “Hey, can we get enough (PCB assembly, metal, parts) for runs big enough that we’re not constantly in backorder.”
Because we had to revisit 17 products this year—in addition to introducing new stuff!
To be clear, no sane person signs up for 5 full redesigns, like we got roped into when the AKM plant burned down, and definitely no engineer who isn’t barking mad wants to do minor-to-moderate changes on 12 other products on top of that.
That’s why I’m tired.
But that’s OK. There are great things ahead. New things, better availability, improved processes.
Yes.
I know.
Stop laughing.
2021: Now With 4X the Bonage!
When we went into 2021, we knew we were boned. The AKM factory had just burned down, and we used AKM DACs and ADCs in 5 of our key products: Modi, Modius, Hel, Fulla, and the AK4490 module used in Jotunheim and Ragnarok.
With AKM being off-line for at least a year, it meant we were looking at 100% redesigns of all of those products.
To add insult to injury, 3 out of 5 of those products had just been redesigned—
to use AKM DACs and ADCs!
Oops.
But yeah. Fulla 3 had already been redesigned to Fulla 4. Hel had become Hel 2. The AK4490 module had become the AK4493/Unison module. All done. Finished. And both Fulla 4 and Hel 2 already had been scheduled for production. They were kitted and ready to go.
Which meant we had a choice: postpone Fulla 4 and Hel 2 for however long it took to get the ESS designs right (maybe 6 months, maybe more), or proceed with Fulla 4 and Hel 2 and then swap to ESS after production had started.
We decided to go ahead and produce Fulla 4 and Hel 2, mainly because (a) they were done, and (b) we had plenty of stock of AKM parts. But let’s not forget (c) the unknown development time for ESS.
Luckily, all went fairly well. We managed to start with AKM, then transition to ESS with only a 2-month or so hiatus. Not great, but not horrible, either. And we still kept enough AKM stock to continue producing Modi and Modius with those parts.
Aside: current runs of Modi and Modius AKM are the last, though…those are going ESS shortly.
But the curveball of “5 new products you didn’t expect, and right now!” was just the start of, well, getting boned. As the subhead says: 2021, now with 4x the bonage!
Here’s what else happened:
1. The parts debacle.
About mid-year, our purchasing partner came to me and said, “You know that ancient regulator you use, the one that’s been around since the earth cooled, the one that literally everyone on the planet has used one time or another, you know the one they used to give out in Cracker Jack boxes as the free prize? Yeah, well we can’t get it reliably anymore.”
Cue my stomach doing backflips. Because the part he was talking about wasn’t just in one or two or three or four products…it was in a TON of stuff. It was in so many things I had to look it up: 12 products.
12 products potentially stopped cold.
Because here’s the thing: everyone uses that old regulator because they are (a) good, and (b) cheap. And with a few pre- or post-processing tricks, they can be (c) exemplary, and still (d) pretty darn cheap.
And here’s the rub: because everyone uses them, there are literally no alternatives.
Sigh.
Aside: this is not why you get into engineering, but if you get into engineering, this is something you’ll have to deal with.
After some conversation with our purchasing and warehousing partner, it became clear that this wasn’t a complete dead-stop oh-crap we’re-really-screwed thing, but a stupid-package thing. As in, we could get all of those silly regulators we wanted—but not in the package size we designed in.
Aside: yes, as in the exact same part in a different case. Sometimes for fun, they’ll even give you different pinouts for different parts suffixes. This causes lots of confusion. Get used to this if you go into engineering too.
Well, that wasn’t so bad, but it wasn’t great, because minor-to-moderate redesigns of 12 products is again, nothing a sane engineer asks for, but it’s significantly less sane to forego producing 12 popular models.
So what did we do? We created a “universal regulator” footprint and dropped it into all of the affected boards.
Sounds easy, right?
Well, in some cases it was, but you still have to run prototypes to make sure you didn’t screw something else up in the process of making a simple change. Which means running boards, stuffing boards, full testing, etc.
Aside: again, if you’re getting into engineering, expect this to happen. Murphy is a dick.
But in the case of 4 of the 12 products, the new regulator footprint ran us into space constraints (as in, it’s a lot larger than the old one). Which meant much more major surgery. Which meant more chances to screw something up in the prototype phase.
Oh, and the final kick in the nuts: all 12 products are mine. None of them are Mike and Dave’s.
So. Yeah. Why am I tired? This is a big reason why.
But that’s not the end.
2. The capacity problem.
We’ve been under-capacity in PCB assembly all year, as you may have noticed.
If you haven’t noticed, this is how you notice: you say, “So where’s my Saga+?” 6 weeks after you ordered it with a 6-8 week leadtime. Or you notice that over half of our products aren’t in stock. Or you see that some of them are estimated to be shipping 10-12 weeks out.
Now, this under-capacity problem was not for lack of trying to expand. We brought two new PCB assemblers on this year, and re-engaged with another. For those of a mathy bent, that’s 4x the PCB assemblers we had going into 2020.
Should be fine, right?
Weeeeeeeellllll…in short, Murphy strikes again. Two of the new partners had much longer ramp-up time than expected (in some cases not their fault, we’ve also had parts shortages to factor in). One of them is much better at the larger, more complex products, but we needed to order the boards to our specs, which took a while to figure out. Two of them are very, very good, but have limited capacity (and we are happy they are honest with what they can take on and what they can’t.) And, at the end, when we thought we were getting things handled, tragedy struck our oldest PCB assembly partner, and they lost a critical person. Lost, as in, he died. Then the other two partners got sidelined with COVID—as in, in the hospital.
Aside: here’s the thing: there are humans behind all these cold things made of metal and glass. Unexpected crap happens. We’re trying really hard. But sometimes things simply aren’t gonna happen.
So has the assembly problem been surmounted? Not entirely. But we’re working on it. And we’re certainly now in a better place.
But the ideal of everything in stock, all the time…lol, nope. Not yet. Sorry.
3. The metal troubles.
PCB assembly would be enough, but we also had metal troubles this year. This was mainly due to one metal supplier—the one we worked with almost from the very beginning of the company—getting bought out and losing interest in our business.
Aside: making things? Get used to this. Especially if they are shiny, consumer-finished things, or if you become perceived as being “difficult.” As soon as your metal vendor can shuffle you off for an industrial contract doing, like, control panels or engine covers, you can end up in a world of hurt.
This was a bit of a surprise, because the metal supplier gave no indication of losing interest—they just stopped shipping, or shipped garbage.
So we moved more production over to our stamping partner. This was fine—until they started running into capacity restraints.
We also gave a bunch of work to a start-up staffed by the dream team of our ex-supplier, the ones who figured out how to do most of our stuff in the first place—but their ramp-up time has been much longer than we expected.
Is the metal situation getting better? Perhaps marginally. I expect we’re still going to have some problems to work out. And I expect silver to continue scarce. Sorry about that, but better to have the truth, right?
4. The tube scarcity.
This has gotta be the most “you gotta be kidding” thing yet: not being able to get enough tubes for Freya+.
I mean, we went with new production tubes for a reason with Freya+—we’d be assured of a good supply! We wouldn’t have to hunt down NOS tubes in large quantities, or sort piles of old Russian tubes! It would be great!
That is, until
you can’t get enough new production tubes.
Sigh. Again.
So what are our options?
- Use different new production tubes. We already are using all that Tung-Sol, Electro-Harmonix, and JJ can supply us. That isn’t enough. Done and done.
- Use super-fancy new production tubes and raise the price. Yeah. Asked about that. They offered to sell us 25 of them. We need like 5000. So no.
- Try to sort decent 6N8S from the Russian stock we have. Unfortunately these all seem to have loose micas. Which means they are microphonic, and it’s not fixable. Sorting appears to be a fools game. Sorry.
- Sell without tubes. Already doing this. Nobody really wants this option, statistically.
- Bring back LISST. Except not many people bought LISST. Still, this may be necessary.
- Redesign for a good, non-microphonic tube we have. Like the 6N1P. This “Freya Noval” needs different operating points, but it might be interesting to some people, because you could use 6922, ECC88, etc in it as well. We’ll see.
Sooooooo….yeah….beginning to see where we are coming from? This is a profoundly weird year. And with a lot of ICs stating out-of-stock status continuing to March 2023 (yes, 2023, not 2022), it may continue to be, well, interesting.
Digesting Texas
“Well, of course you’re tired!” some of you are going to cry. “You decided, in the middle of all of this nuttiness, to start up a whole new facility in Texas!”
Yes, we did.
And actually, it’s been relatively pain-free.
Now, Alex may argue with me a bit. He would certainly argue with me back in February and March of 2021, the months of The Big Freeze and the insanity of start-up. To be fair, we did ask him to basically replicate what he did in Valencia over 8 years…in about 8 weeks!
But, as of now, I think we’ve achieved a relative miracle: we opened new manufacturing in a new state, found a good balance between the two facilities, and still managed to grow the business.
A lot of this has to be credited to Alex and Tyler on the Texas side, and Elvis and Amy on the California side. Alex and Tyler both chose to relocate to Texas. Elvis and Amy are picking up where they left off in California.
Over the course of the year, Texas absorbed literally every wall-wart or USB-powered product that Schiit makes. As of now, if it has a wall-wart or runs off USB, it comes from Texas. If it has an AC power cord, it comes from California.
This split works really well, because California doesn’t have to deal with the bulk and complexity of wall-warts, and Texas doesn’t have to deal with AC power and burn-in racks. California can concentrate on the more complex products—the Freyas, the Loki Maxes, the Ragnaroks…and, yes, Jotunheim and Asgard as well. Texas can concentrate on the products we need to make efficiently in large volumes. It’s a nice split, and it appears to be working very well.
And just like that: Texas, digested.
What We Did Right
After the positive note above, let’s continue riffing on what we did right. Because we did a ton of things right this year! It’s easy to forget that, in the ongoing weirdness.
So what did we do right?
We managed the challenges well. I think this is really the biggest thing. Yeah, we had a ton of not-good stuff happened, but we managed the challenges as well as we could. Nobody stuck their head in the sand, nobody ran from them, nobody checked out. And we kept things kinda-sorta available, and even improved backorders thru the year. At a time where it’s still hard to get a car, a game console, or the exact computer you want, that’s really the main thing.
- Good transition to ESS. We responded to the news of the AKM factory fire early, beginning redesign for ESS immediately, getting production orders placed, and prioritizing the 5 products we needed to change. As a result, the ESS transition has been essentially a non-issue, and we are only now running out of AKM…almost exactly as planned.
- Better communication. We invested in better ways to communicate with you, more people to run the comms, and changed the ways we defined backorders to make it much more clear how long you might be waiting. This, in turn, allowed us to throttle orders more effectively, reducing the deepest of the backorders.
- New systems. As of right now, we’re on the verge of deploying a real ERP system, which will dramatically increase the visibility of what we have and what we need on an ongoing basis. Until now, we’ve been doing this with meetings and spreadsheets. The time is long past for that.
We introduced, or re-introduced, a reasonable amount of new products. As in, we didn’t go completely insane with product intros, but nor did we completely fail at getting things out. For those of you who missed it, here’s what we did
- Loki Mini+. The first product this year was a replacement for the Loki Mini, based on the improved gain stage we debuted in Magni 3+. Why change this product? For better performance.
- Hel 2/2E. Hel 2 came about due to us wanting to add both USB-C and optical inputs, as well as UAC1/2 autoswitching—making it a much more flexible product. Hel 2E came about when we switched to ESS DACs.
- Fulla 4/E. Fulla 4 came about because we wanted to Hel-2-ify it, and Fulla E happened, again, due to the ESS changeover.
- Lokius. The first expansion of our equalizer line, came about almost as a lark, an affordable 6-band, LC or gyrator-capacitor topology EQ. The final port of call before expensive options set in.
- Yggdrasil Flavors. The first DAC upgrades in quite a while surprised a lot of people—mainly due to our refusal to brand one the One True Path. But our decision to introduce “flavors” was based on lots of development and listening—fully 10 different revs of the Yggdrasil were built and evaluated!
- Loki Max. The big-boy EQ, the only one of our crazy products I had hoped to have in 2021…maybe this is the year of the EQ? Or the Year of Ongoing Sacrilegious Products? Who would have thought we would have three equalizers—in 2021?
- Freya S and Saga S. And we re-introduced the Freya S and Saga S. This was mainly a move to use up metal we found in the warehouse, but the popularity of Freya S, and the ongoing tube scarcity, means it remains in the line.
So, in total, 8 products. Or 10 if you count the Fulla 4/E and Hel 2/2E as separate products. Which you should, since they were ground-up redesigns. Only the chassis were the same. In the old days, that would be a heck of a lot of products. Now, not so much!
We found the right balance between California and Texas. As mentioned above in “Digesting Texas,” we managed to get manufacturing set up almost without incident, and split the line almost seamlessly, and logically, between the states. Sometimes, things do go right.
What We Did Wrong
So what did we do wrong? Well, there was some pain in getting Texas and California set up on equal footing…but, in short, not a lot.
I mean, seriously, everybody stepped up this year. If it wasn’t for Elvis reworking the way we did production in Valencia, our backorders would have been monumentally worse. If it wasn’t for Amy kicking ass on purchasing for Valencia, again, we’d have been in deep trouble. If Alex hadn’t set up Corpus with such efficiency, and if Tyler hadn’t managed the bilocation of the company seamlessly, it could have been disastrous. If Laura hadn’t stepped up in shipping and customer service in Valencia, things would be even crazier, and if Scott hadn’t done the same thing in Corpus, it would be completely nuts. If Asa and his crew in Valencia, and Evan and his team in Corpus hadn’t stepped up to test and qualify more products than ever, we would have been hamstrung. If Marv hadn’t created a whole new visibility system for our backorder status, and if Lisa and Jen hadn’t managed it, things could be very different.
Bottom line, it was a wild year. Why beat ourselves up?
Let’s leave it at that.
Can It Get More Bizarre?
After a bonkers 2020 and a somehow-even-weirder 2021, I guess the logical question is “can it get any weirder?”
Last year I would have said, “Absolutely not!” Now, I’ll just say, “Hell if I know!”
I mean, the parts situation could get worse. We don’t really know. That would put a big damper on everything. Or our metal and PCB assembly headaches could continue. Again, we don’t know. (Though we are actively working to make sure the metal and PCB problems are non-problems.)
To cut to the chase: I certainly hope it isn’t as bonkers, but I wouldn’t be terribly surprised if the weather continued weird, with transdimensional showers.
And to be a bit more positive, you should be able to count on at least three things from us next year, because we already have metal and parts and stuff for these:
- Tyr. AKA “the big amp.” We’re just waiting for production on this one. Long delayed due to the PCB debacle.
- Folkvangr. AKA “the crazy 10-tube headphone amp.” Just waiting for some metal. We have tons of tubes for this one!
- Urd. Just waiting for Dave and Ivana to pronounce it done.
“Ah, come on,” you say. “That’s it? Nothing more?”
Go back and read this whole chapter again.
There. Is. Nothing. Certain.
However, I can say this: our experience with bringing hardware equalizers back from the dead and in forging a new, saner, “flavorful” path to the future (rather than all-march-in-lockstep-to-the-One-True-Upgrade) has whetted our appetite to skewer some more sacred cows. So expect more sacrilegious barbecue in the future.
Aside: provided we can get the parts!
Aaaand…some of the things we’re taking on are so entrenched, so ingrained, that, if we do it right, you’ll shake your head and say, “Wait a sec. Have we been doing this wrong for years…or for decades?”
Beyond that, not a word.
Other than I’m looking forward to 2022. And I hope you are too!
