What Went Wrong With Dyson’s $500 Toothbrush? Manufacturing Lessons for Hardware Companies

WEBVTT
00:00:03.714 --> 00:00:08.085
You're listening to china manufacturing decoded from the Sofeast group
00:00:13.107 --> 00:00:18.657
hi welcome back to china manufacturing decoded it's episode 349 today adrian
00:00:18.657 --> 00:00:21.513
hosting here and renault's on board with me hi renault
00:00:22.233 --> 00:00:23.766
Adrian how are you doing.
00:00:24.432 --> 00:00:25.245
Yeah good how are you
00:00:25.849 --> 00:00:29.424
All good We made it back to China,
00:00:31.414 --> 00:00:38.545
after the long break here, actually it was a combined Mid Autumn Festival plus,
00:00:39.177 --> 00:00:42.672
October 1st called the golden week, a kind of double holiday.
00:00:44.790 --> 00:00:47.541
Now things are getting started again finally.
00:00:48.396 --> 00:00:52.286
I got you on board today because we were talking about this before the podcast
00:00:52.286 --> 00:00:56.376
right there was it's quite an interesting story i think that anybody who's developing
00:00:56.376 --> 00:01:00.016
or manufacturing a hardware product can probably learn something from
00:01:00.996 --> 00:01:07.006
the start of last month september 2026 dyson who are probably best known for
00:01:07.006 --> 00:01:09.713
vacuum cleaners they launched this
00:01:10.596 --> 00:01:17.346
really newfangled toothbrush called the camera jet which is just the 500 renault and
00:01:18.256 --> 00:01:22.676
that's something that took them six years to develop hundreds of engineers and
00:01:22.676 --> 00:01:28.026
it's got a lot of quite sophisticated technology on board so if you're listening
00:01:28.026 --> 00:01:31.386
you can probably guess what's coming now within days of the launch
00:01:32.054 --> 00:01:36.791
people were reporting that the product was dying maybe they never really worked
00:01:37.255 --> 00:01:41.812
and then they quickly took it offline and it wasn't available
00:01:42.462 --> 00:01:49.016
so what could cause this and and we're going to dig into this i will make the point now that as
00:01:50.016 --> 00:01:55.546
It appears to be back on sale so i think that the issue was found and maybe
00:01:55.546 --> 00:02:00.656
solved but that doesn't mean that there's nothing that we can we can learn and
00:02:00.656 --> 00:02:03.389
so i think there are some sort of quite familiar
00:02:04.132 --> 00:02:10.801
hardware development risks that we can talk about so this was supposed to be waterproof
00:02:11.479 --> 00:02:15.706
you know what the what are the critical processes in mass production that that
00:02:15.706 --> 00:02:20.116
need to be controlled here you mentioned validating production so
00:02:20.916 --> 00:02:24.919
how do you validate it properly before a major launch was it in this case perhaps
00:02:25.343 --> 00:02:29.576
how do you react to early field failures there's a lot of stuff here So this
00:02:29.576 --> 00:02:35.186
isn't only for Dyson or toothbrushes, but I think it can apply to a lot of hardware consumer products.
00:02:35.186 --> 00:02:39.356
So when you saw this story, because you mentioned this to me,
00:02:39.716 --> 00:02:42.699
I was aware of it, but actually it was a good idea to talk about it.
00:02:42.966 --> 00:02:45.967
What stood out to you from a manufacturing perspective?
00:02:47.186 --> 00:02:51.406
Very painful i i can only imagine right so they,
00:02:52.114 --> 00:02:58.982
they spent years doing the product development they fine-tuned it they say okay this really you know,
00:02:59.562 --> 00:03:04.882
there's really some kind of benefit that we can talk about then they they tool
00:03:04.882 --> 00:03:07.930
up they they industrialize the product,
00:03:09.004 --> 00:03:12.442
but they need to keep bit kind of secret, right?
00:03:12.442 --> 00:03:17.422
It's a little bit like Apple, because they want a big bang of,
00:03:17.882 --> 00:03:23.358
oh, this is new, a new product, look at that. Oh, Dyson is coming up with something crazy again, right?
00:03:25.329 --> 00:03:34.012
They built their image on nicely designed and nicely made products that you
00:03:34.012 --> 00:03:35.638
want to have in your home, basically.
00:03:37.042 --> 00:03:43.630
So, vacuum some cleaners, lots of hair dryers and things around the hair.
00:03:44.713 --> 00:03:48.632
What else? They launched something that was kind of like a mask and at the same
00:03:48.632 --> 00:03:51.844
time, like over the head headset.
00:03:52.163 --> 00:03:55.652
I don't think that would work, but at least, you know, people are like,
00:03:55.652 --> 00:03:58.048
okay, wow, you know, this is differentiated, right?
00:03:58.396 --> 00:04:02.762
This is a, if you really have that problem, you really need to have clean air
00:04:02.762 --> 00:04:05.362
and so on, like, you know, sort of be in your bubble.
00:04:05.902 --> 00:04:08.860
This is, you know, might be a great product, okay?
00:04:10.114 --> 00:04:13.323
They wanted to go into cars, and then at the last moment, they kind of pulled
00:04:13.323 --> 00:04:17.881
the plug. Okay, I think they did well. This got way too competitive.
00:04:18.523 --> 00:04:22.523
And now, so they launched something in the mouth, right? I think it was the
00:04:22.523 --> 00:04:24.749
first time Dyson does something for the mouth.
00:04:27.516 --> 00:04:31.683
And obviously, it's a very dynamic category. There's been a lot of electric
00:04:31.683 --> 00:04:37.112
toothbrushes that came up with extra features.
00:04:37.361 --> 00:04:42.893
So why now and not 10 years ago? I think there's a lot of sensors and a lot
00:04:42.893 --> 00:04:46.243
of things you can do, a lot of cool things you can perceive.
00:04:46.771 --> 00:04:51.198
And the idea here, if I understand, is that what they call the camera jet,
00:04:52.140 --> 00:04:55.343
right, is that basically there's a camera that picks up, hey,
00:04:55.343 --> 00:04:58.823
here's there's a little bit of a space between two teeth or something.
00:04:59.482 --> 00:05:05.241
Okay, I'm going to send a jet of water to do like flossing, kind of.
00:05:05.815 --> 00:05:09.043
So you don't need to do something weird around your teeth, you know,
00:05:09.443 --> 00:05:12.813
flossing, and you can do much more than a toothbrush.
00:05:13.039 --> 00:05:15.734
So, okay, pretty cool.
00:05:16.549 --> 00:05:22.070
The thing is, yeah, going from vacuum cleaners and hair dryers to electric toothbrush,
00:05:22.991 --> 00:05:27.106
that, you know, there's some new challenges, let's say, right?
00:05:27.425 --> 00:05:30.769
And from pushing air to pushing water, right?
00:05:31.757 --> 00:05:38.153
And from like fighting dust to fighting saliva and all kinds of water and stuff
00:05:38.153 --> 00:05:40.183
like, you know, there's some new challenges.
00:05:40.183 --> 00:05:46.803
So it's, I'm sure they had some kind of risk analysis. I'm sure they thought
00:05:46.803 --> 00:05:50.699
about all this, but apparently they get surprised, right? Right.
00:05:51.457 --> 00:05:58.563
And they did that also, I'm pretty sure they did that outside of China because
00:05:58.563 --> 00:06:03.433
they take it as a point of pride to make everything outside of China.
00:06:03.433 --> 00:06:06.387
So that might be made in Malaysia or someplace like this.
00:06:07.982 --> 00:06:13.015
Which is fine. There's a lot of good expertise in Malaysia for electronics.
00:06:13.561 --> 00:06:21.297
However, if you're going for the most experienced manufacturers of electric
00:06:21.297 --> 00:06:24.937
toothbrushes, you go to China, right?
00:06:24.937 --> 00:06:30.939
So anyway, I'm not saying this has any relationship with their issues, just a side note.
00:06:31.817 --> 00:06:36.436
So yeah, to answer your question, very painful, very painful. You do all that work.
00:06:37.401 --> 00:06:43.221
So you build up sort of the surprise, then you unveil it to the world,
00:06:43.784 --> 00:06:49.357
and you're very proud of it, and okay, there's some press, and within, what, a few weeks?
00:06:49.653 --> 00:06:55.255
You need to say, ooh, we stop, we withdraw from the market, stop everything,
00:06:56.039 --> 00:07:00.107
just one production batch, but, you know, it's just one production batch,
00:07:00.107 --> 00:07:04.264
why would they have to remove, to stop everything, right?
00:07:05.093 --> 00:07:08.327
Any issues with traceability here? Because if you have good traceability,
00:07:08.327 --> 00:07:11.747
if it's really one production batch, and you confirm it's one production batch,
00:07:11.747 --> 00:07:14.264
then maybe you don't stop everything?
00:07:15.257 --> 00:07:17.927
Or maybe they had no inventory and it was the only one they were setting,
00:07:17.927 --> 00:07:20.811
I have no idea, right? I don't want to jump to conclusions.
00:07:21.527 --> 00:07:24.337
Or maybe they were not sure yet it was only one production batch.
00:07:25.357 --> 00:07:28.002
So anyway, very painful because...
00:07:29.776 --> 00:07:37.211
I'm sure it killed all the momentum of the launch right and that translates to,
00:07:38.194 --> 00:07:42.312
a loss maybe tens of millions of dollars of free advertisement you know that,
00:07:42.872 --> 00:07:47.221
and suddenly becomes uh negative publicity that's pretty bad.
00:07:48.393 --> 00:07:52.695
Yeah yeah it launched on the first and i think failures were being reported by
00:07:53.258 --> 00:07:56.385
no later than the ninth so about a week and
00:07:56.838 --> 00:08:01.222
then in the middle of the month it was it was taken off sale for a period of
00:08:01.222 --> 00:08:05.292
time so yeah all within two weeks you're absolutely right it doesn't mean the
00:08:05.292 --> 00:08:09.152
product doesn't work now but yeah the the launch was overshadowed very much
00:08:09.152 --> 00:08:11.192
by poor review because reviewers from
00:08:11.612 --> 00:08:15.571
you know the publications like i don't know wired and other other ones like that
00:08:16.193 --> 00:08:20.382
they received these products tested them and they've said oh that's dead i will
00:08:20.382 --> 00:08:24.272
some of them i didn't even work was never able to brush my teeth it's not it's
00:08:24.272 --> 00:08:28.862
not Not the publicity that a famous company like Dyson wants when they do a new launch.
00:08:28.862 --> 00:08:34.894
So, yeah, I mean, imagine if you're a sort of a newer business and this happens, it's really bad.
00:08:35.846 --> 00:08:39.770
Right. And that's weird. I mean, you're talking about an out-of-the-box failure.
00:08:41.082 --> 00:08:47.042
But that's something that looks like a manufacturing issue. Yeah, definitely. Right.
00:08:48.042 --> 00:08:51.716
But how come it was, I mean, I'm pretty sure they were tested at the end of
00:08:51.716 --> 00:08:56.917
the line. What happened? How come they arrived and they're dead out of the box?
00:08:57.336 --> 00:08:59.446
This is pretty weird.
00:09:00.178 --> 00:09:05.266
If I were the product manager at Dyson or top management, I would really be
00:09:05.266 --> 00:09:09.193
asking very tough questions right now. This is just weird.
00:09:09.692 --> 00:09:12.566
From some of the articles I read, I think some of the journalists asked for
00:09:12.566 --> 00:09:16.861
a follow-up on the returned products because they sent them new ones.
00:09:17.413 --> 00:09:22.746
And i don't think they've given them a follow-up as of yet so i don't know whether
00:09:22.746 --> 00:09:24.906
they're still trying to work it out i think one of the
00:09:25.566 --> 00:09:29.746
one of the big issues that was reported was potentially the waterproofing was
00:09:29.746 --> 00:09:34.316
problematic right because this is a device that's definitely going to be very
00:09:34.316 --> 00:09:38.046
wet it makes itself wet because it's spraying out like
00:09:38.466 --> 00:09:40.742
a mouthwash type solution so
00:09:41.294 --> 00:09:45.106
it lives in a wet environment it's getting wet it's wet in your mouth and it's
00:09:45.106 --> 00:09:50.206
spraying water out that's actually stored inside in a reservoir as well so the
00:09:50.206 --> 00:09:52.966
whole thing is really wet and yeah so
00:09:54.006 --> 00:10:00.096
dyson claimed that it's ipx7 water resistance right and that may well be true
00:10:00.096 --> 00:10:04.818
because of course it's been tested extensively but i read somewhere that um
00:10:05.546 --> 00:10:09.596
one of the journalists had spoken to someone and they said well this this is
00:10:09.596 --> 00:10:14.356
fine when you're testing the product and validating it but in the process of
00:10:14.356 --> 00:10:19.091
shipping when products electronic products often undergo you know big
00:10:19.806 --> 00:10:25.126
variations in heat and humidity that can actually affect the cure on waterproofing
00:10:25.126 --> 00:10:30.986
whatever's been used for it right so they suggested that this could be an issue
00:10:30.986 --> 00:10:33.301
that had caused the devices to basically
00:10:34.326 --> 00:10:39.266
not work almost straight away this is some supposition Yeah,
00:10:39.266 --> 00:10:42.253
Exactly. We should take that with a grain of salt because,
00:10:43.057 --> 00:10:48.090
People who have most of the data are the manufacturer, right? Dyson.
00:10:49.739 --> 00:10:53.802
I'm sure they get their hands on some of these failed units as soon as they could.
00:10:54.283 --> 00:10:58.190
They run a failure analysis, try to understand what happened, right?
00:10:58.469 --> 00:11:06.089
So, but if they don't communicate, then we have to play the armchair analysts
00:11:06.089 --> 00:11:08.415
here. Yeah. And okay, it's fun.
00:11:08.740 --> 00:11:14.099
So, basically, yes, as I said earlier, they went from products that basically
00:11:14.099 --> 00:11:19.910
blow air to products that contain and blow water.
00:11:22.023 --> 00:11:26.462
I would not be surprised if, for example, yes, the environmental and reliability
00:11:26.462 --> 00:11:33.092
test protocols were not really a good fit, which is what that person,
00:11:33.092 --> 00:11:35.000
that commentator suggests.
00:11:35.407 --> 00:11:43.022
Yeah. Now, ask Dyson how they test, you know, how they validate the reliability
00:11:43.022 --> 00:11:46.123
of a new vacuum cleaner, you know, or hair dryer.
00:11:46.361 --> 00:11:52.092
I'm sure they're very good. You know, they are probably among the very best
00:11:52.092 --> 00:11:53.982
in the world at doing that, right?
00:11:53.982 --> 00:11:57.892
So when they have new product design, you know, they're not going to release
00:11:57.892 --> 00:12:02.754
something onto the market that will die prematurely.
00:12:03.108 --> 00:12:06.932
You know, they will have done a lot of smart testing. They will really try to
00:12:06.932 --> 00:12:13.302
reproduce the usage conditions and the stresses and all these kind of things.
00:12:13.302 --> 00:12:19.152
But now something that goes to the mouth, I'm sure they hired people from some
00:12:19.152 --> 00:12:21.643
of the best companies doing that.
00:12:22.067 --> 00:12:28.424
But still, you know, as an organization, you need to burn your own database of issues.
00:12:28.696 --> 00:12:31.541
And this is part of your organizational knowledge, right?
00:12:32.458 --> 00:12:35.892
And I would not be surprised if this is really missing. Now,
00:12:35.892 --> 00:12:39.238
this is being burnt right now, as we're speaking, right?
00:12:39.693 --> 00:12:43.592
When you go into a new category, that's always the kind of things that can happen.
00:12:44.212 --> 00:12:50.282
And again, if you make that in, I don't know, for example, Malaysia rather than
00:12:50.282 --> 00:12:52.930
China, then also you cannot count as much,
00:12:53.400 --> 00:12:57.602
on a really good manufacturer that's been doing that for a long time and themselves
00:12:57.602 --> 00:13:01.649
have their own opinion and experience on that, right?
00:13:02.432 --> 00:13:06.930
That might play a role, again, with just armchair analysts here,
00:13:07.732 --> 00:13:09.490
a bit removed from the action.
00:13:11.180 --> 00:13:15.793
But waterproofing, when you have a product like this, that's really a beast.
00:13:16.147 --> 00:13:24.262
And I believe that most companies that launch an electric toothbrush,
00:13:24.976 --> 00:13:29.975
and it's the first time they make this kind of product, have serious issues.
00:13:29.975 --> 00:13:32.621
Most of them have serious issues with waterproofing.
00:13:33.225 --> 00:13:38.862
Like, I've seen that, but of course I haven't run a wide survey.
00:13:39.297 --> 00:13:44.874
But I'm pretty sure it's extremely common, okay? I'm sure it happens at least 70% of the time.
00:13:45.355 --> 00:13:48.804
It's really an electric toothbrush that's designed from scratch.
00:13:49.134 --> 00:13:52.176
At least 70% of the time, if it's the first time to do it.
00:13:53.037 --> 00:13:56.863
Why? Because this thing,
00:13:57.445 --> 00:14:02.655
you hold in your hand, put in your mouth, is constantly fluids dripping,
00:14:02.928 --> 00:14:07.395
various kinds of fluids from clear water to whatever, saliva,
00:14:07.395 --> 00:14:10.972
to, you know, something with toothpaste, all kinds of things like that, right?
00:14:11.472 --> 00:14:16.195
It kind of drips onto the product, so it's like the entire outside shell of
00:14:16.195 --> 00:14:20.330
the product has to be waterproof.
00:14:21.245 --> 00:14:26.291
And plus there's like the inside, you know, the circuit of the water and,
00:14:26.552 --> 00:14:33.145
you know, the motor, all these kind of things, the PCB board, you know.
00:14:33.165 --> 00:14:42.150
Now, if there's a little bit of leakage over time, it can cause complete failure of the product, right?
00:14:42.824 --> 00:14:50.556
So that's quite common. And it needs to be designed, and that's really the key,
00:14:50.874 --> 00:14:54.205
is that the whole system has to be designed the right way.
00:14:54.725 --> 00:14:57.655
We're approaching this, right? The way you design the enclosure,
00:14:57.655 --> 00:15:02.200
whatever, the seals, or just the way the different parts mate together.
00:15:03.643 --> 00:15:04.134
You know,
00:15:05.929 --> 00:15:10.356
the tolerances, the materials you pick, how they're going to behave,
00:15:11.489 --> 00:15:17.028
including, as you were mentioning, if you have weak changes of temperature,
00:15:17.272 --> 00:15:21.434
is one of them going to expand or retract faster than the other?
00:15:21.684 --> 00:15:23.530
And after a while, what does that mean?
00:15:24.469 --> 00:15:27.862
For the boot product assembly, and so on and so forth. Okay,
00:15:28.037 --> 00:15:32.324
so you have to design the product.
00:15:34.443 --> 00:15:38.893
Basically, you need a lot of lessons learned here, right? You can't just take
00:15:38.893 --> 00:15:42.903
a mechanical engineer out of school. Hey, you know, this is something cool.
00:15:42.903 --> 00:15:44.323
You know, put this together.
00:15:44.549 --> 00:15:46.279
You're going to have all kinds of problems.
00:15:46.593 --> 00:15:49.838
Okay. You really need a team that has gone through this.
00:15:51.623 --> 00:16:00.703
And on top of that, you need the manufacturing to be done the right way.
00:16:00.703 --> 00:16:04.295
Okay. So you design the product. You also design the manufacturing. process.
00:16:06.082 --> 00:16:08.636
If there's a little bit of misalignment here or there.
00:16:10.903 --> 00:16:14.556
It's a disaster. So again, I'm not surprised.
00:16:15.402 --> 00:16:22.544
I would have put that as one of the very top, very highest types of risks for this product.
00:16:23.323 --> 00:16:26.544
But again, yeah, in hindsight, everything is easy, right?
00:16:26.939 --> 00:16:31.316
Maybe they did everything right and they were not lucky or something.
00:16:31.671 --> 00:16:37.143
But that's the thing when you innovate and you go first to market with a product
00:16:37.143 --> 00:16:43.855
that's the first of its kind, you're going to discover issues, right?
00:16:44.232 --> 00:16:50.593
You're going to discover it. And you could do like Apple and make hundreds and
00:16:50.593 --> 00:16:56.846
hundreds and thousands and thousands of units and bows and things like that, right?
00:16:57.590 --> 00:17:03.934
But not every company can do that. I'm sure Dyson cannot do that to the same extent as Apple.
00:17:04.294 --> 00:17:07.446
And Dyson cannot rely on,
00:17:09.015 --> 00:17:14.283
Honestly, the very best suppliers in the field, including component suppliers,
00:17:14.283 --> 00:17:18.036
etc., etc., module suppliers, and so on,
00:17:20.163 --> 00:17:23.823
the way Apple does. Nowhere near, right? So,
00:17:26.430 --> 00:17:34.342
in that regard, Dyson is closer to average SME, right? That's what I would suspect.
00:17:35.880 --> 00:17:42.293
Okay. So if we move away from the Dyson a bit, right, because this is a good,
00:17:42.293 --> 00:17:48.630
it's a good starting point, but you spoke about waterproof products and the risks there, right?
00:17:48.926 --> 00:17:53.783
Even if we know that the waterproof design is sound, could it be the manufacturing
00:17:53.783 --> 00:17:55.734
process that then undermines that?
00:17:56.580 --> 00:17:58.819
Oh, yeah. Yeah, of course. So.
00:17:59.962 --> 00:18:06.986
You could have a great design and make some pilots in Unity,
00:18:06.986 --> 00:18:13.136
DVT, and really subject them to all the IP67.
00:18:15.356 --> 00:18:21.199
Lab, you know, they get jets and so on. It's rather punishing already.
00:18:22.136 --> 00:18:24.785
And pass that, right? And simulate whatever.
00:18:26.536 --> 00:18:31.206
Two times a day, times two years, you know, or times, I don't know,
00:18:31.206 --> 00:18:32.815
maybe two minutes in the mouth.
00:18:33.157 --> 00:18:36.866
Simulate the life of the product. So, okay, and come out and say,
00:18:36.866 --> 00:18:40.019
okay, this is a great product. The product design is great.
00:18:41.203 --> 00:18:45.835
Then you send it to manufacturing, and it's not that great, right?
00:18:46.364 --> 00:18:50.702
Well, why? Why would it be not that great?
00:18:51.271 --> 00:18:56.986
Well, the manufacturing process itself might not be mature enough.
00:18:56.986 --> 00:19:00.025
Maybe you did not give them enough time to do things right.
00:19:00.493 --> 00:19:05.976
Maybe you did not really engage them from the beginning of the engineering validation
00:19:05.976 --> 00:19:08.156
runs and design validation runs, etc.
00:19:09.101 --> 00:19:17.695
With a sufficient number of units made with the final manufacturing process.
00:19:18.100 --> 00:19:22.354
Maybe you switch from one factory to another and the new one kind of,
00:19:22.528 --> 00:19:25.883
you know, had to figure out a lot of things on the fly very fast, right?
00:19:26.295 --> 00:19:30.138
I don't know the case here, but it's always a possibility.
00:19:30.649 --> 00:19:34.610
So don't always blame the manufacturer, right?
00:19:35.732 --> 00:19:42.236
Maybe also the test, sort of end-of-line testing or sub-assembly testing is.
00:19:43.662 --> 00:19:46.992
Not very mature, not very good at picking up the issues.
00:19:47.422 --> 00:19:52.263
That might be something that is quite common.
00:19:52.762 --> 00:19:56.280
Everybody wants to go to market fast and the test,
00:19:57.107 --> 00:20:03.821
The test benches and everything are not ready yet, right? Now for electric toothbrushes,
00:20:04.396 --> 00:20:11.646
I'm sure you go on Adibaba and you find kinds of test fixtures that you can buy off the shelf.
00:20:12.087 --> 00:20:18.350
But here we're talking about a very new kind of product that does things in a new way.
00:20:18.663 --> 00:20:22.773
So I'm sure they had to build a lot of their own test fixtures, right?
00:20:23.023 --> 00:20:31.368
So again, don't blame the manufacturer necessarily. these issues might be self-inflicted by Dyson.
00:20:34.093 --> 00:20:38.913
Or a company that manages the whole industrialization and supply chain for Dyson.
00:20:38.913 --> 00:20:41.967
Again, I don't know. I don't know how they're managing that.
00:20:43.493 --> 00:20:49.113
Also, it could be poor process control. So maybe when you do some pilots and
00:20:49.113 --> 00:20:54.227
you have engineers looking over and fixing every little issue and everything,
00:20:54.396 --> 00:20:58.116
okay, everything turns out really good and the tests,
00:20:59.364 --> 00:21:05.693
the test stations catch the few non-conforming products and you're like,
00:21:05.693 --> 00:21:08.881
okay, yeah, we can ramp up the volume, no problem.
00:21:09.375 --> 00:21:12.747
And then when it's really just managed by the production people,
00:21:13.095 --> 00:21:17.424
then maybe the process is no longer really in control.
00:21:18.689 --> 00:21:22.773
The way they, I don't know, apply the glue, maybe, I don't know,
00:21:22.773 --> 00:21:26.917
just a little bit of drift in the glue machines or whatever. Right?
00:21:27.299 --> 00:21:33.787
It could be a lot of things. Maybe simple, very dumb kind of issues. Maybe the.
00:21:36.052 --> 00:21:42.693
Screwdriver torque maybe is not calibrated for a few days. You know, it kind of drifts.
00:21:43.406 --> 00:21:48.921
Maybe there's cleanliness issue. maybe it could be a lot of things, right?
00:21:49.768 --> 00:21:56.185
And you have one of these issues for, you know, a few hours,
00:21:56.185 --> 00:22:03.530
maybe one shift, or a few days, and then suddenly it can lead to total failure of the product.
00:22:04.551 --> 00:22:11.088
And then everybody goes into panic mode, and then, oh, we withdraw it from the market, sorry.
00:22:11.482 --> 00:22:14.415
There's a problem with at least one production batch. we still need more time
00:22:14.415 --> 00:22:20.080
to think about it, to find really the cause and the root cause of the issue.
00:22:20.509 --> 00:22:26.405
But for now, we stop selling because we don't want to keep delivering 40 products,
00:22:27.045 --> 00:22:30.655
to the market and then have to manage your returns and everything,
00:22:30.655 --> 00:22:32.245
which is much more expensive, right?
00:22:32.745 --> 00:22:36.416
So that happens, that happens.
00:22:36.781 --> 00:22:45.576
And it's much more likely to happen toward the beginning of mass production than six months later.
00:22:46.865 --> 00:22:52.484
Because you do things on a small scale on one or two production lines,
00:22:53.599 --> 00:22:59.895
then you start to ramp up and you don't have, again, as I mentioned,
00:22:59.895 --> 00:23:04.365
all the Quakes engineers and the R&D engineers, everybody also from the Dyson
00:23:04.365 --> 00:23:06.656
side, you know, looking at everything.
00:23:07.703 --> 00:23:11.066
Easy to have someone make a mistake. Very easy.
00:23:11.396 --> 00:23:15.656
That's the problem. Now, six months later, everybody knows, oh,
00:23:15.656 --> 00:23:17.986
we've got to be careful about this, we've got to be careful about that,
00:23:17.986 --> 00:23:22.166
the work instructions are more mature, the way they patrol on the line and when
00:23:22.166 --> 00:23:25.716
they check, and so on, the test fixtures, etc.
00:23:26.296 --> 00:23:29.586
I mean, everything is more mature, much more likely to catch the issues.
00:23:29.586 --> 00:23:35.391
But at the very beginning, that's where it's likely, let's say.
00:23:35.391 --> 00:23:38.456
And I think this is a really good illustration of that because literally this
00:23:38.456 --> 00:23:42.636
is a product that's just gone into mass production and launched so that it kind
00:23:42.636 --> 00:23:46.225
of backs up what you're saying with regard to
00:23:46.840 --> 00:23:51.796
whether it was a batch and they were being cautious by taking it off sale the
00:23:51.796 --> 00:23:56.536
only evidence I've got from Dyson was that they said that and I quote one production
00:23:56.536 --> 00:24:00.700
batch of jet tubes had experienced a process issue right
00:24:01.176 --> 00:24:07.396
so the jet tubes are I think the spread like the part of the assembly that sprays the
00:24:07.876 --> 00:24:12.429
mouthwash-type solution to clean the plaque from the gaps of teeth.
00:24:13.793 --> 00:24:18.426
So one production batch suggests that they probably did have some sort of traceability,
00:24:18.426 --> 00:24:21.178
of course, and they took everything to be safe.
00:24:21.614 --> 00:24:26.843
The only thing is, if it's a defective part to do with the spraying,
00:24:27.453 --> 00:24:28.796
I don't know whether that backs up.
00:24:29.316 --> 00:24:34.826
It could do but like the device is basically dying straight away because that
00:24:34.826 --> 00:24:39.296
to me sounds like water ingress uh which i don't know whether that would be
00:24:39.296 --> 00:24:43.823
so related to that that particular part of the of the device but
00:24:44.310 --> 00:24:48.226
but yeah i think they basically had an issue with a batch and then took everything
00:24:48.226 --> 00:24:51.094
off to be cautious that's that's what happened yeah i mean if it's
00:24:52.676 --> 00:24:57.416
If the product is just taken out of the box, people don't even put water in
00:24:57.416 --> 00:24:59.047
it, and it's not turning on.
00:24:59.645 --> 00:25:03.466
There's probably some other issue, because on the production line,
00:25:03.466 --> 00:25:07.403
I don't think they're testing it with water, not out of it anyway.
00:25:11.557 --> 00:25:16.217
I don't know if they were DOA or whether the people started filling them up
00:25:16.217 --> 00:25:19.807
and tried to turn it on and then basically they just stopped working straight
00:25:19.807 --> 00:25:21.525
away. I think it's more the latter.
00:25:23.371 --> 00:25:24.915
Yeah, it might be.
00:25:25.704 --> 00:25:29.957
So you were talking about some
00:25:29.957 --> 00:25:33.717
of the issues that can go wrong as we get towards the mass production.
00:25:33.885 --> 00:25:39.981
And this kind of ties back into having a structured MPI process, right?
00:25:40.934 --> 00:25:42.604
Yeah, very much so.
00:25:44.080 --> 00:25:49.676
You develop a complex, ambitious project like this, obviously with a big budget,
00:25:50.233 --> 00:25:56.034
and you target relatively high volumes, you get to go.
00:25:59.694 --> 00:26:05.575
With a relatively serious and nicely resourced NPI process.
00:26:06.724 --> 00:26:11.274
So you work with a manufacturer that knows what they're doing and can put a
00:26:11.274 --> 00:26:17.944
lot of engineering resources, on this, but also on your side,
00:26:17.944 --> 00:26:20.184
you try to guide things, right?
00:26:20.701 --> 00:26:27.004
You try, even when the product is still in development, you already try to get
00:26:27.004 --> 00:26:34.938
to make some builds, even with runs of 20, 50 pieces or even more units, I should say,
00:26:35.642 --> 00:26:38.674
and try to see, okay, it's breaking here, it's breaking here,
00:26:38.694 --> 00:26:42.016
it's breaking here, it's breaking here, we have all these issues. that's fine.
00:26:42.538 --> 00:26:47.094
Now we're in a much better position in the next run, in the next sort of design
00:26:47.094 --> 00:26:52.024
iteration, to do it right. And then again we try to force it through all the
00:26:52.024 --> 00:26:55.487
way to manufacturing the parts.
00:26:56.674 --> 00:27:01.441
At some point you start to make it maybe with soft tooling much earlier than you would otherwise.
00:27:01.714 --> 00:27:08.041
And you assemble, you test, you pack, you do extra testing for reliability etc.
00:27:08.337 --> 00:27:15.821
And then you kind of see all the issues that you have, but by doing this repeatedly,
00:27:16.216 --> 00:27:18.108
three, four, five times.
00:27:19.464 --> 00:27:24.004
You develop the muscle everywhere, including at the manufacturer's place,
00:27:24.503 --> 00:27:28.754
about, okay, this is exactly how we do this, and this is exactly how we check
00:27:28.754 --> 00:27:30.649
quality, and this is exactly how we test,
00:27:31.794 --> 00:27:34.834
the products, and this is how we do this, and this is how we do that.
00:27:38.374 --> 00:27:39.303
And you really,
00:27:40.439 --> 00:27:49.650
try to think of the risks and you try to put the finger on the highest sources
00:27:49.650 --> 00:27:54.268
of risk and you try to implement corrective actions for all of them, right?
00:27:54.559 --> 00:27:57.860
Because, again, you have time. You're starting to do all this and you see all
00:27:57.860 --> 00:27:59.539
the things kind of breaking and.
00:28:00.600 --> 00:28:03.820
You audit the component suppliers, you audit this, you audit that,
00:28:03.820 --> 00:28:09.190
you keep looking for issues, and then you start to have a long to-do list and
00:28:09.190 --> 00:28:11.633
then a lot of people have to do a lot of the work, right?
00:28:12.081 --> 00:28:17.660
But that's how you get to the point where by the end of production validation,
00:28:18.320 --> 00:28:25.650
we'll call free production, you are already ready to make the product.
00:28:25.650 --> 00:28:29.770
Now, not a very high first pass yield, right? Meaning that some of the products
00:28:29.770 --> 00:28:32.030
still have to be reworked and so on.
00:28:32.308 --> 00:28:36.430
Okay, but it's getting better and better. you're already at a point where you
00:28:36.430 --> 00:28:42.299
can go ahead and start to sell products and have a plan for ramping the quantity up.
00:28:43.060 --> 00:28:46.850
If you don't do all this, say, okay, let's just make some prototypes,
00:28:46.850 --> 00:28:51.687
like on a bench, without really involving the manufacturing partner,
00:28:51.919 --> 00:28:53.812
without really doing all the testing.
00:28:54.090 --> 00:28:59.280
Okay, I see this issue, this issue, you know, let's just go as an engineering team.
00:28:59.280 --> 00:29:03.172
Let's just do the next iteration, next iteration. Okay, now we throw it over the wall,
00:29:03.718 --> 00:29:07.410
to a company that will start to make the mold because now we're in a hurry,
00:29:07.410 --> 00:29:12.370
but we need to start to get the mold done, and then the PCBA supplier,
00:29:12.370 --> 00:29:16.303
and we get all this with the different parts, and we put it together.
00:29:17.320 --> 00:29:20.063
And we engage a contract manufacturer that would make it.
00:29:20.860 --> 00:29:28.470
That's the wrong way, right? You need to have the manufacturer do reviews and
00:29:28.470 --> 00:29:30.183
tell you what the issues are, right?
00:29:30.566 --> 00:29:36.280
What we call design for manufacturing, for testability, for quality, etc.
00:29:37.860 --> 00:29:42.010
But also people call it DFX, design for excellence, because it's not just designed
00:29:42.010 --> 00:29:44.855
for manufacturing and assembly. It's much more than that.
00:29:46.041 --> 00:29:51.496
If you don't integrate the feedback from a good manufacturer,
00:29:52.198 --> 00:29:55.732
you're going to have all kinds of issues when you industrialize.
00:29:57.766 --> 00:30:03.345
And we don't know exactly what process Dyson has followed in this particular
00:30:03.345 --> 00:30:08.996
case, but because of the nature of their product, right, which is an innovative product, which
00:30:10.155 --> 00:30:14.137
involves many, many patents and a lot of, you know, engineering,
00:30:14.747 --> 00:30:19.013
it's probably you would argue a riskier proposition right
00:30:19.530 --> 00:30:25.275
so they've gone for what looks like a big bang release first of september on
00:30:25.275 --> 00:30:31.767
sale in many many markets right which then compounds the problem if people start getting failures
00:30:32.231 --> 00:30:36.165
you're just going to have a lot more in one big go which then as we see with
00:30:36.165 --> 00:30:40.358
the with the bad press and all the rest of it gives you more of a problem
00:30:40.788 --> 00:30:45.865
would you say that for a product like this in your opinion maybe like a phased
00:30:45.865 --> 00:30:51.144
release would be a bit less risky so maybe just testing one market first
00:30:52.312 --> 00:30:59.934
So it's always going to be less risky if you go more gradually to market, right?
00:31:00.520 --> 00:31:06.637
But then you lose a lot of the free press, you know, free publicity.
00:31:07.635 --> 00:31:13.184
And that, in the case of a company like Dyson or Apple, right,
00:31:13.607 --> 00:31:20.308
they have, I'm going to say, there's a lot of journalists that want to cover the latest gadgets.
00:31:20.308 --> 00:31:25.416
Yeah, and talk about it, right, including in leading publications.
00:31:25.735 --> 00:31:34.408
So when you are at that point, missing that free publicity is extremely expensive, right?
00:31:34.843 --> 00:31:38.388
Or missing a part of it. So you say, well, we're just going to start to sell
00:31:38.388 --> 00:31:42.278
it just in the UK, and that's it.
00:31:42.278 --> 00:31:46.438
Then the newspapers in the US are much less likely to cover it.
00:31:46.438 --> 00:31:50.788
And then when you say, oh, we start to sell it in the U.S., well,
00:31:50.788 --> 00:31:54.148
it's much less of a novelty. So, you know, again, you're not really going to
00:31:54.148 --> 00:31:58.901
get covered. So I can understand why they want to do that.
00:31:59.116 --> 00:32:06.197
And what Apple does in this case is that they will study the processes to death.
00:32:06.958 --> 00:32:10.808
You know, they would make so many parts and study them, you know,
00:32:10.828 --> 00:32:16.581
or see statistically, do we have a problem here, or is the process stable, and so on.
00:32:17.283 --> 00:32:23.788
Let's make a lot of units, and let's subject them to a lot of testing,
00:32:23.788 --> 00:32:31.487
and do this and do that. And they basically have a nearly unlimited budget, right?
00:32:32.415 --> 00:32:40.618
And time also is less important than quality, right? So, they have an unlimited
00:32:40.618 --> 00:32:44.601
budget of money and time, basically, to get to where they want.
00:32:44.996 --> 00:32:48.928
And they don't release something in beta, right? They will release something
00:32:48.928 --> 00:32:55.228
that has to be up there in terms of quality, reliability, etc. So, they have to...
00:32:56.844 --> 00:33:00.451
They have to put in a lot of money and a lot of time, right?
00:33:00.602 --> 00:33:06.436
Now, Dyson, from what we see, puts a lot of time or so, and they probably put a lot of money.
00:33:07.166 --> 00:33:08.862
But you still have the risk, right?
00:33:09.704 --> 00:33:16.394
And if you're Apple and you release a new phone, well, you have so much expertise
00:33:16.394 --> 00:33:20.271
in what can go wrong on a phone and how people exactly use the phone and everything,
00:33:21.535 --> 00:33:26.599
you're much more likely to prevent these kind of issues Then if you're Dyson,
00:33:27.017 --> 00:33:31.510
you go from basically air blowers to water blowers, right?
00:33:31.891 --> 00:33:40.458
That is a big, big change. It really is about waterproofing issue. I'm not surprised.
00:33:42.144 --> 00:33:46.214
But this is sort of a big company or big brand, I should say,
00:33:46.214 --> 00:33:48.022
big brand problem, right?
00:33:48.823 --> 00:33:52.654
A lot of people listening are not in this case. They're not going to get free
00:33:52.654 --> 00:33:55.607
publicity just because they have something out there, right?
00:33:55.961 --> 00:34:01.664
They're not going to be in all the top publications. So in a lot of cases,
00:34:01.664 --> 00:34:04.111
what we suggest is don't do that. Don't rush.
00:34:04.835 --> 00:34:08.454
Sort of said, you know, get some real users, not your family,
00:34:08.454 --> 00:34:13.623
not your friends, to use it, give you feedback after one or two months.
00:34:14.302 --> 00:34:17.561
And that's how you prevent this kind of issue.
00:34:18.159 --> 00:34:22.344
And don't make it very visible, like on a Kickstarter or something like that,
00:34:22.344 --> 00:34:24.998
because then a lot of companies in China will see it also.
00:34:25.154 --> 00:34:29.108
If you don't want that, do everything offline at the beginning, right?
00:34:29.514 --> 00:34:33.324
When you're really ready to push and put a lot of fuel in it and everything,
00:34:33.324 --> 00:34:37.484
then you can do a lot of things online and have a lot of visibility,
00:34:37.484 --> 00:34:44.299
which brings you more customers, but also maybe more competitors, right? But until then,
00:34:45.584 --> 00:34:51.112
don't do everything online and don't be worried about talking to real users right,
00:34:51.838 --> 00:34:56.604
i mean it's not like they have a cousin within shenzhen hungry to copy you right
00:34:56.888 --> 00:35:02.555
just you know get your products in in people's hands real people right so,
00:35:03.417 --> 00:35:09.969
that's the advice we keep telling companies that sell a new product but they don't always listen.
00:35:10.670 --> 00:35:15.017
Yeah yeah one other thing i've thought about is that no doubt because we're
00:35:15.017 --> 00:35:19.767
coming up into the christmas period now that's why they've launched in the september
00:35:20.056 --> 00:35:24.817
and i imagine that if they didn't launch at this time of the year
00:35:25.297 --> 00:35:29.960
it would have been pushed back to next year and you don't want to miss the christmas period so
00:35:30.529 --> 00:35:36.735
so that that that time scale doesn't surprise me for dyson but yeah if you're
00:35:37.036 --> 00:35:40.057
if you're a smaller sme maybe the
00:35:40.857 --> 00:35:45.736
the not pushing so hard is is definitely sensible i can understand that
00:35:46.717 --> 00:35:51.887
can we finish off just by talking a little bit about when we get products in
00:35:51.887 --> 00:35:57.397
the field finally and when we start to get reports back from customers hopefully
00:35:57.717 --> 00:36:01.212
not too many negative ones right but what's what
00:36:02.217 --> 00:36:07.357
what would you do to manage sort of field failures what sort of process would
00:36:07.357 --> 00:36:09.082
you suggest can be followed.
00:36:10.052 --> 00:36:14.168
Yeah, there is a process about that. I'm pretty sure they have something about that, but,
00:36:14.881 --> 00:36:21.029
you try to gather a lot of information and fail samples as fast as possible,
00:36:21.522 --> 00:36:26.316
and send all of that to a central point that will be able,
00:36:27.156 --> 00:36:30.076
to classify it and see if there are patterns.
00:36:32.657 --> 00:36:37.823
And then it's about classifying it, doing a paratoanalysis, and working on,
00:36:38.428 --> 00:36:44.616
the top one, two, three types of issues, fitting this back to product design
00:36:44.616 --> 00:36:47.976
and manufacturing processes and whatever.
00:36:48.643 --> 00:36:50.976
Maybe it's a component, maybe it's whatever.
00:36:51.748 --> 00:36:56.085
Maybe a tube or whatever it is in this case.
00:36:57.503 --> 00:37:04.916
And trying to find ways to improve. But you You should not focus,
00:37:06.434 --> 00:37:13.056
excessively on getting the products and managing the returns to keep customers
00:37:13.056 --> 00:37:14.022
happy, that's one thing.
00:37:14.643 --> 00:37:20.426
But here, it's very important to get the learning and to push all of that again
00:37:20.426 --> 00:37:23.284
in a central location or in central software somewhere.
00:37:23.807 --> 00:37:27.064
Everything categorized the same way, analyzed the same way.
00:37:27.958 --> 00:37:33.882
And based on that, okay, what is the next step? Do we have a big problem? Do we have a problem?
00:37:35.232 --> 00:37:38.845
What can we do now to reduce the risk, right?
00:37:39.477 --> 00:37:45.142
Yeah. And I guess we have to accept that there's never going to be zero issues
00:37:45.142 --> 00:37:47.342
when we bring a new product to market, right?
00:37:47.342 --> 00:37:53.329
But as you've said, find those issues quickly, find them early,
00:37:53.816 --> 00:37:59.502
and take action quickly to prevent it from becoming a massive crisis,
00:37:59.502 --> 00:38:03.162
which appears to be what Dyson did with this particular batch.
00:38:03.358 --> 00:38:09.062
Because as I say, they're back on sale now. So one assumes that Dyson are confident
00:38:09.062 --> 00:38:13.282
that that particular issue is now not an issue if they're back on sale.
00:38:13.282 --> 00:38:17.769
So they've clearly taken quite quick action with that.
00:38:18.077 --> 00:38:22.882
With regards to the products, I mean, we've discussed the difficulties of making waterproof products.
00:38:22.882 --> 00:38:26.952
So that's probably a takeaway for the listener if you're thinking of going into
00:38:26.952 --> 00:38:31.432
that particular category or an electronic consumer product that needs to be
00:38:31.432 --> 00:38:34.478
waterproof, even if it's not an oral care product.
00:38:34.745 --> 00:38:39.032
Yeah, the risks involved there. Maybe you might need a longer development time
00:38:39.032 --> 00:38:44.138
and more validation before going to market than a product that doesn't need to be waterproof.
00:38:44.294 --> 00:38:47.203
So that's food for thought, I think, as we leave the listener.
00:38:48.522 --> 00:38:51.142
Yeah, I agree. All right. All the first to add.
00:38:51.922 --> 00:38:55.483
So yeah if you're listening let us know if you bought one of these dyson
00:38:56.014 --> 00:39:01.092
toothbrushes or camera jets and um what what were your experiences with it we
00:39:01.092 --> 00:39:03.642
don't have one yet maybe somebody will buy me one for christmas
00:39:04.102 --> 00:39:08.852
i don't know right but but yeah be interesting it's certainly an innovative
00:39:08.852 --> 00:39:10.082
and interesting product and
00:39:10.522 --> 00:39:15.019
and yeah they're going into new territory here so i suppose some teething problems
00:39:15.385 --> 00:39:19.460
are to be expected with products like this so that's what we've been focusing on today thanks for now
00:39:20.729 --> 00:39:23.893
All right thanks that was an interesting case i think.
00:39:24.687 --> 00:39:25.906
Yeah that's it so
00:39:26.306 --> 00:39:31.206
When we talk about um a prog lounge usually that's because there are issues.
00:39:32.142 --> 00:39:37.202
Yeah well unfortunately yes unfortunately but at least it's a learning experience
00:39:37.202 --> 00:39:41.673
so there's always something we can take away from it and learn so thanks dyson for that
00:39:42.842 --> 00:39:45.482
Right all right let's limit that then.
00:39:46.282 --> 00:39:52.192
Right okay uh we'll be back next week. Thanks again for listening to this podcast
00:39:52.192 --> 00:39:53.913
brought to you by the Sofeast Group.
00:39:54.429 --> 00:39:58.214
We're on a mission to provide you with everything you need to manufacture effectively
00:39:58.353 --> 00:40:00.670
in Asia, including inspections,
00:40:01.070 --> 00:40:04.751
auditing, new product development support, contract manufacturing,
00:40:04.977 --> 00:40:10.347
3PL warehousing and fulfillment, and much, much more across Asia's key manufacturing areas.
00:40:10.735 --> 00:40:17.794
Visit us at sophast.com, that's S-O-F-E-A-S-T dot com to learn more and get help.
00:40:18.357 --> 00:40:22.282
If you've enjoyed the podcast today, please do rate, review and share because
00:40:22.282 --> 00:40:24.191
it will really help others discover us too.