Feeling more constrained
Industry constraints are continuing to get worse — now what?
Speaker: Terry West
read podcast TRANSCRIPT:
I just read the latest Jabil Commodity Intelligence Report and it's a fascinating read. It's about an eight-page ten page PDF download that they publish quarterly. Jabil one of the largest contract manufacturers of electronics equipment in the world. They publish this quarterly report that talks about how hard it is to get parts electronic parts and what's happening with pricing those parts. They published their first one in 2017 and they updated it for Q1 2018.
I was just going over it again today. A lot of things in the last week have come together around constraints in the industry which promoted yet another podcast (like I tried to do every quarter) on this problem. It's getting quite a bit worse and we knew that (my prior podcast forecasted that we would see industry constraints lasting through 2000 anything certainly into 19).
And now we have more data points, not just the Jabil report but also some data points from my manufacturing teams in China as well as here domestically. We have one particular manufacturer in China who we asked if they would be able to purchase some components on our behalf in China. Resistors, if you know if you don't know what that is, that's a little tiny component of which we typically use 50 or so on every single board. They're used everywhere in the electronics industry. They typically cost $0.01 a piece.
The manufacturers came back to us and said, "we can't buy resistors right now." I know that for some of you, it may not have the impact that it has certainly had on me (if you don't you live in that world). For a manufacturer to not be able to buy resistors is a mind-boggling statement. Yageo, who's one of the largest manufacturers of resistors in the world (based in China I believe), basically said to this manufacturer of ours "don't even bother putting orders in. We won't be able to fulfill them." There are billions and billions and billions of resistors used every year in the process of manufacturing electronic equipment. So, for Yageo was one of the largest to tell the manufacturer "don't even bother putting an order in" is mind-boggling. They also said, "we have a get-well plan that gets us healthy by 2020. So you'll be able to start putting orders again in 2020."
To me, that's mind-boggling, that basically there will be vast vast areas of the industry that will effectively be 'lines down' and shut down, potentially for almost two years.
This could certainly affect you if you have products that use electronics. There are certain suppliers that have guaranteed supply. People like Apple (building iPhones) and some of these manufacturers building (whether it's an Amazon Echo Dot or whatever it is). And if you're not one of those privileged, very small few who are also struggling to get the capacity, you're going to have a problem.
Now we at Serious obviously take this very seriously, because we want to be able to meet commitments with customers. I've been encouraging customers now for two quarters to backlog orders going out for the next six months on a rolling basis. And I appreciate those customers if you're listening, that you've done that. So we have many customers, in fact, most of our customers, have placed orders (scheduled orders, monthly or weekly drops) all the way for the next six months. I had one customer looking to place orders right now (here we are in May of 2018) and they're looking to place orders for Q1 of 2019. When we have those orders on the books, we can go proactively buy inventory, find where we need to go do whatever we need to do to find inventory, for those builds. I'm happy to say so far that, although were slightly delayed on some meeting some customer needs, at this point in time we've met all of our customer commitments. And that's a big deal.
There's a lot of vendors out there that are not meeting customer commitments because they can't get parts. That's a big problem. What that means downstream is customers can't build (whatever it is they're trying to build with the machines, the devices, the products) and they're going to get lines down and it's going to be a difficult situation to navigate.
There's really no one at fault here. There's no one to blame. You know I'm pointing my finger at my part suppliers, sure. But the parts suppliers, whether they're chip vendors, resistor manufacturers, capacitors, coils, wire, or whatever it is -- there's only so much equipment they have. There's only so much capacity they have. Since 2000 when the bubble burst, there hasn't been a lot of capital invested in new manufacturing plants (certainly domestically here in the U.S.). I would venture to say there's very little new manufacturing capacity if anything we've lost manufacturing capacity over the past eight years. As a result, we're not prepared for an economic boom. Which is kind of the mode we're in right now - a fairly dramatic economic recovery in the US and somewhat globally as well. Which is driving demand for sophisticated electronics and many new products or are things out there are becoming electronic.
The Jabil report points to a couple of things. One, electric vehicles are really starting to take off globally and they have significantly higher electronics components in electronics in the electric vehicle versus the pure gas powered vehicle. Even a gas vehicle has something like 30 or 40 processors and its support circuitry. But an electric vehicle doubles and triples that. And so when you multiply the economies of scale that go with automobiles globally, that's a vast increase in the complexity of a car. The number components that are purchased.
The second thing they mention of course is continuing sophistication of smartphones and tablets that are driving larger usage of the factories that build chips primarily.
The third thing that is driving capacity needs is IoT. The proliferation of connected devices and things that used to be non-connected becoming connected. Whether it's your Ring doorbell or your Fitbit, you as a human being now are connected through your smartwatch, Fitbit, or whatever it is. All kinds of things that were not connected before that used a modest amount of electronics are now becoming connected now. Therefore, their utility is increasing so not just getting the communication capabilities you need yet more processing more memory. Once you've got those things connected, there's a natural tendency to increase the value of those things which have more features, which adds more electronics, and more software. You get this value increase very quickly across an enormous number of different IoT devices. Not all just little tiny sensors stuck on poles to measure wind speed, temperature pressure, gas flow. or other, but more sophisticated devices.
In fact, one of our new products that we're launching this year the SCM318 is an amazingly sophisticated piece of technology. I've been in the industry since the mid-80s. And when I talked about a hardware board to do something (whether it was a control machine or run a device) in the old days that would be a processor, a fuel little chips to talk the outside world component, a little bit of support circuitry, and a processor. The SCM3 team for us is an off-the-shelf platform to basically be the "brain board" in many connected devices. Whether it's a Bluetooth router, ice cream machine, a piece of equipment for the production line, or a kiosk for ordering pizza (whatever it is) the SCM 318 is designed to be the brain board for all of these types of devices. If I compare the complexity of that board with that 1980s platform, we now see processing power from 8MHz and 4MHz megahertz back in the 80s to 400 MHz now on these processors. We see memory which used to be 256 bytes of memory to up to 256 megabytes of memory (a million times more memory than we had in the 80s). The surround circuitry of power supplies and Wi-Fi, Bluetooth, Ethernet, and all those things that didn't exist on many of the boards back in the 80s that and our just table stakes that are expected as part of these new boards. And you look at the SCM318 that has all that stuff on board ready to go.
And I know that Moore's Law says that every 18 months we can double the processing power double the number of transistors on a chip. But in spite of Moore's Law, the demand is significantly outpacing the supply and with limited capital over the past 18 years to invest in new factories the world has really been caught flat-footed. And is trying to struggle to meet this demand.
As a result, if you're a purchasing manager or a strategic buyer out there today, whether you're using serious products or any other products that are based on electronics, I would suggest to you that if you don't have your backlog in place for the next six to eight months with every one of your suppliers that has electronic content. You may be buying a box as an OEM subsystem that goes inside a pump or whatever it is that goes inside your machine, or you may be buying any of my modules from us, or communications control module like the SCM318; or you may be buying chips and building your own PCs to mix in with other pieces like a little off the shelf printer model that you're buying or a scanner model that you're buying. All of those things contain electronics. As challenging as it is, you need to as a supply chain team look out over the next rolling six to eight-month period and put a baseline buy plan in place with all of your suppliers for that period. And every month add another month. Review it, modify it, and add another month the back end. And in an ideal world, you've got six to eight months of orders on the books with every single thing that you buy that includes electronics.
That doesn't guarantee you're going to get it. But one thing I can say and guarantee is that you are far more likely to get the product if you had an order on the books than if you say, " I don't have to worry about this. I don't have a problem right now. My supplier tells me there's no problem." And if the supplier tells you they're no problem they're lying. This is a huge industry problem right now. The right answer from every supplier that you have as it should be, "yes it's a huge industry problem. Yes, we're keeping our customers that have orders on the books up to the best of our ability. And yes, if you put backlog in place we'll do everything we can to try and meet your needs." That is the only answer that people who sell electronics products can give to you. And the other answer is an exaggeration.
I'm absolutely emphatic about this and that's the answer I give to customers right now. When they asked me what my lead times are I say, " well if you wanted to order 10,000 pieces of this today I might be able to dribble a few out for you over the next several months based on excess capacity. But the reality is I'm probably 26 to 30 weeks to be able to fulfill that order." Why? Because processors are 26 weeks, memory is 26 weeks ( if you can get it), and resistors we've got people like Yageo saying don't even bother ordering until 2020.
So the dependencies of putting together electronic products and then assembling them, testing them, and building them (right now) are typical, for, at random, new, big order are typically 26 to 30 weeks. That's an amazing amount of time, six to eight months, Place an order now, get your product and six to eight months. Which is why when you have backlog on the books when you have orders on the books with all your suppliers they can plan and say, "hey, I'm going to prioritize people giving me orders and give me visibility over people who come in out of the blue and say, " I know I didn't want to give you any forecast or anything. And now any product." You go, "I'm sorry, but I've only got this much product to ship. I'm going to prioritize the customers that have orders on the books over customers and needs that haven't got orders on the books."
I know this seems fairly straightforward. I've been through this scenario this industry problem about four times now over my career. Many of you may have also gone through these things and you know that it's always darkest before all of a sudden everything turns. That tipping point between supply and demand. I know that when we got into this product this constrained situation we had about 2 - 3 weeks notice and things started getting progressively worse. And the antenna for people who have been through this process in the past start to get raised and think, "oh, I feel it I feel it's happening" and they take actions very quickly. Sometime in 2017, I think it was around Q3 when we when my antenna went up (I'm sure many of you yours did as well). We saw, within a period of 3 - 4 weeks memory prices doubled, or tripled, within three weeks.
And now we're seeing the ripple effects. I would say that it's probably just as quick on the other side when supply, all of a sudden, exceeds demand and it tips the other way. What happens, especially in the commodity business, is that buyers place multiple orders at different suppliers because they can't get product, so they're out there desperately placing orders on varieties of places for the same product. A hundred here, a hundred there, a hundred there, a thousand here, thousand there, and thousand there. The reality is they may place the order in three different places with a thousand pieces and then when they get the thousand pieces, they cancel the other two orders. If they get it. What happens is that we end up with this overbought situation where the demand is actually not what the industry really need because everybody is desperate for products and all the bookings that happen. When supply starts to free up everybody cancels the multiple bookings very quickly. Because if you only need a thousand, you don't want to end up with all three of your suppliers shipping you 1,000 and you're stuck with 3,000. So that's minute this thing turns, you as a supply chain person start canceling those multiple orders. Then, all of a sudden everybody thought they had 3x demand, that they actually didn't, the whole thing collapses, everything turns. All of a sudden, everybody's swimming and product prices fall through the floor. We're not there yet.
We need to manage this very closely across the industry, certainly at Serious and our customers, so that we don't carry too much inventory at any given time. Which is why, quite frankly, I don't have a huge capacity for upsides and many manufacturers don't. Because we can't afford to go buy memory at $5/chip and then charge you downstream for it even more, because of markup and management the constraints, and your prices go up. If we all end up with too much inventory and everything collapses, all of a sudden that five dollar memory chip is a buck. Then we'd all be sitting on inventory that we pay too much for and there's no other option except take a loss and write it down to what it's currently worth. So none of us want to be overextended in terms of inventory. At the same time, we all want to get our products and ship them.
So again, hopefully, this insight is useful to you as a supply chain team. The 'magic bullet' here, to the best of our abilities, for us as Serious to our suppliers, and certainly you to Serious into all of your electronics suppliers, you have to put backlog on for the next six to nine months. And if you don't, the likelihood of you being 'lines down' because you can't get one of your pieces for six months becomes increasingly high. That's certainly not what I want to see for your business.
If you have any insights, data points, or contradictory data, reach out to me. I'd be absolutely happy to hear that and reflect that on a future podcast and er and revise my view of the world based on data. But you know this is what I'm saying. And I wanted to share it with you quickly.
The challenges of HMI and IoT can be overwhelming. Listen weekly to our Captain – Terry West, CEO – on industry trends he sees and implications for the embedded systems designer.
You'll find practical thought leadership conversations covering all facets of embedded technology, human machine interface evolution, architectural challenges, IoT readiness, and much more.