ZPM: The Kickstarter Failure That Revolutionised Espresso

James Hoffmann
3 followers
May 1, 2023
Today, we're gonna talk about this, the ZPM Nocturne.
Now, this was a Kickstarter project.
It had 1,500 backers and not a single unit was delivered.
I didn't think it existed,
but it turns out one of my Patreon supporters
had one of the beta-tested units and he got in touch
and said, did I want to buy it?
And yes, yes, I did.
Now, the story of ZPM is interesting,
but what's more interesting to me
is that you can draw a line from the failure of ZPM
through to the creation of what I think
is probably the most influential espresso machines
in the last decade,
a machine that's kind of revolutionized
our understanding of espresso, the Decent DE1.
Now, before you ask, yes, it does work,
and yes, we are gonna make some espresso with it,
but let me give you the backstory on ZPM.
The Kickstarter campaign launched back
on December 11th, 2011, and this was back in the day
when videos were kind of charming
and not necessarily big budget affairs.
Their plan was to use a bunch of off-the-shelf components
combined with a couple of their own bespoke pieces,
I think their own Thermoblock design and group head design,
to create a very technically capable espresso machine
that was very, very cheap.
Their proposed pricing for Kickstarter
was $200 for this unit, which was a ridiculous price.
Now, the planned retail would be $400,
but either way, for the kind of features they were offering,
this was a very, very compelling proposition.
When the Kickstarter finally closed,
they had raised $370,000.
Now, in terms of modern Kickstarters,
that's not that much money, but at the time,
this was the sixth highest funded Kickstarter
in its history.
Turns out their pricing was perhaps not sustainable.
Something we do talk about on this channel
is the challenge of manufacturers
scaling up what they're doing,
and I think people often underestimate
how difficult it is to go from
some kind of fully functional prototypes,
which is essentially what this is,
through to mass manufacturing thousands of units.
I don't think they had the skill necessarily.
I don't think they had the money necessarily to do it.
Now, I'm not trying to pick over the bones of it.
I just think they were very passionate founders
who were perhaps in way over their heads.
The project didn't fail quickly.
In fact, it was October 2015,
nearly four years after launching,
that they posted their final update to Kickstarter.
And when they had shared the fact
that the company had ultimately failed
and they would not be shipping units,
the response from their customer base,
from their community, was angry.
So angry, in fact, that it generated enough interest
for the New York Times to come and write a piece
about ZPM and its community.
Now, if you read the article,
and it is linked in the description down below,
towards the end of the piece,
you're introduced to a new character in all of this,
that of John Buckman.
Now, John, you may know as the founder,
the creator of the Decent Espresso Machine.
And in the piece, he's in talks with ZPM
to kind of buy the remnants of the company,
seeing as John is in London right now,
I thought I would just go and ask him in person
about how his interactions and experience with ZPM
maybe influenced his creation of the DE1.
So firstly, I saw the ZPM video
and I felt like it was gonna be quite hard for me
to figure out how to make espresso,
as I had no guide.
And when I saw this video of these two nerdy guys
and they were both technologists,
and I'm a technologist,
and I thought I'm gonna trust them
to be my guides into espresso.
And they told me what grinder to buy,
which was great, because I didn't know that either.
While they were falling apart,
my businesses were falling apart too.
I had a music business and a book swapping business,
and that was all cratered,
and I was looking for something else to do.
They, one day, sent an email out saying,
we can't ship, we've run out of money, we're really sorry.
At which point, I happen to be in California,
I give them a call, I say, hello.
It sounds like you need a little bit of money,
because they said, we're so close to the finish line,
we just need another million or something.
I don't remember what the number was.
And I thought, okay, well, I can write that check
and take over and deal with adult supervision.
They brought the ZPM over with blue bottle beans,
and they made me quite good espresso.
And so there it was, working, making good coffee.
The next thing I did was call up a friend in Seattle,
because I happen to know Lamar Zoco and Slayer
and Sinezo are up in Seattle.
And Bill Crosland, who'd been involved with the GS3,
was one of the people who was willing to talk to me.
Now, he makes a machine called the CC1.
And what Bill said is, let me show you, kid,
what the economics are of espresso.
And he laid it out to me, and he said,
there's just no way you're gonna make a machine
better than mine at a cheaper price.
What if I just sell direct?
Then could I spend more money to make a better machine?
And he said, if you're willing to sell direct only,
yes, you can make a better machine.
So Bill essentially sent me in that direction.
And he also said, don't buy ZPM,
because it'll never get safety certification.
And that was just him, his background,
it's, you're just gonna die on that altar.
And in fact, they were dying on that altar.
That's what they were talking about,
was the safety certification was never gonna happen.
But the real birth of Decent Espresso is finding that,
I think there's a market for what I call at the time,
Seattle, Italy.
So Italian espresso, which is what I liked at the time,
but with a sort of Seattle spin of high tech app,
that sort of thing.
And that would hopefully reduce the complexity
of making good espresso.
And then the second part,
which was the launch of Decent Espresso,
is everyone telling me, for the amount you will spend
to make ZPM come to market,
you can start from scratch
with people who've made espresso machines before,
and you won't have any of their problems.
Of course I had new problems,
but that's why I went in a new direction.
Because I decided to start from scratch
and make my own machine,
the ZPM guy said, hey, wait a second,
we've shown you everything.
You need to cut us in here.
And they're absolutely right.
I saw the schematics, the software, I saw everything,
and that was influencing the decisions I was making.
So it took a while,
but we got to a point where they licensed
their intellectual property to us,
so we could choose to use it if we wanted to,
in exchange for some money,
and most importantly, a coupon to all the backers of ZPM
that they could apply in the next three years.
It took me about nine months to fall out of love
with the approach that ZPM were taking,
because I tried the same approach,
which is use cheap components
and a bit of software and algorithms,
PID dust, magic fairy dust, PID will solve everything.
That didn't work at all.
My ambition for a machine that I would sell for 1,000 US,
well, we had no product.
And so what I thought I could make for 1,000 US
using essentially two Nespresso machines built together
would not make good coffee.
So this became now a multi-year R&D project.
And it turned out that heating on demand
is incredibly inaccurate on Espresso.
And so we had to come up with new technologies
to deal with that.
What I realized was that we had no idea
how to make Espresso good with light roast beans.
That was actually the problem I wanted to solve.
I had this very strong feeling
that the entire Espresso industry was pre-scientific.
They don't measure anything.
They don't create testable hypotheses.
And testable hypotheses and measurement
are the basis of science.
And we had no idea what pressure, flow, and temperature
would make delicious light roasted coffee.
So let's just make sure that it can do anything.
And in the computer world, we talk about Turing machines,
which is a computer that can run any software.
The idea behind the Decent was to make
a Turing-compliant Espresso machine.
So an Espresso machine that could do any pressure,
any flow, any temperature,
because we have no idea what's gonna work.
In so doing, it became quite expensive to make.
And became a much harder project.
But it also became much more interesting.
So what ZPM really did for me is get me interested,
get my team together, and look at cheap components
and see what we could do with cheap components
to make good Espresso.
And it turns out that if you spend millions of dollars
and develop all this intellectual property,
you can actually take the inexpensive components
and get them to do your bidding.
And so that is how we go from the ZPM to the DE1.
Very different approaches,
but an interesting connecting point between the two.
And I cannot understate
just how influential this thing has been.
By letting us track everything, measure everything,
our understanding of Espresso has fundamentally changed.
How we see flow, how we think about channeling,
all of that stuff has been revolutionized by the DE1.
And I think it's fair to call it
the most influential machine in a decade.
But this is not a review of the DE1.
That will happen eventually on this channel.
This is a video about the ZPM.
It's finally time to fire this thing up
and see what it's capable of doing.
Before we do that, though,
there is a short ad for this video's sponsor,
which is Surfshark.
Surfshark is a fast, easy-to-use VPN service
that is jam-packed full of features
and lets you use unlimited devices with just one account.
And this means that when someone sends me a YouTube video
that's really interesting,
but I click it and it says,
no, this is not available in your country,
I'm not tearing my hair out.
I'm just two clicks and a few seconds away
from being in whatever geolocation I need to be in
to watch the video.
And it's not just YouTube.
This unlocks the streaming services like Netflix
that lets me watch content that isn't available in the UK
that is available in other places.
I want to get all the value I can
out of the streaming services that I pay for,
and this is a great way to do it.
More than this, it gives you
an additional layer of security and privacy
when you're online, whether you're at home
or you're traveling out at a cafe or anywhere else.
As I said, it's unlimited devices with just one account,
and so they have apps for just about everything,
Mac, Linux, Windows, iOS, Android.
You are going to be covered, and I think that is amazing.
If you want to check it out,
you can click the link down below
and go to surfshark.deals.jameshoffman
and use promo code James Hoffman
for 83% off and three months extra for free.
Thank you to Surfshark for sponsoring this video.
Let's switch it on.
Every time I turn it on, my heart is in my mouth
because I'm terrified I'm just gonna break this thing
and then it'll be gone.
You know what I mean?
I don't know if I could repair this.
Someone could, but I couldn't.
Anyway, it's on, it's heating up.
Little temperature display here.
It rockets up in temperature, which is kind of interesting.
Just got a little thermal block to heat in there.
Now, this is obviously a US machine.
It's running on 110.
I've got a chunky transformer under the bench here
to allow me to run this thing on UK power,
but while it's heating up,
we'll do a quick tour of the machine.
It's pretty simple, but there's a couple of notable things
on this particular unit.
So obviously you've got your display on the front here.
It's not a bad display considering its age
and the price point.
It's not horrific.
It's not touch, that's okay.
You've got these control buttons here
and this is something I'm definitely terrified of breaking.
They just feel really flimsy and soft
and I have to be very delicate here.
Two sturdier buttons, one that says start
and one that says mode.
So because it's a thermal block,
if you want to steam milk,
then you'd have to change the temperature
of the sort of thermal block to get steam.
So it can't do both at once.
That's kind of it on the front panel here,
little drip tray, pretty nice.
Now you can see a water tank here.
Actually, if I turn this to the side,
you can see that the water tank does not fit
because this would have been a custom part
that would have cost a lot of money to make.
They didn't get that far.
So this is just, I assume some other machines water tank
that sort of fits well enough, but it works.
Also worth noting here,
you appear to have a mini USB and an SD card slot
and impressively,
we're actually up to temperature on the front.
So we've got a little ready symbol at the top,
which says 95, 94.9 degrees Celsius.
Lovely to see it in Celsius.
Didn't even put that on myself,
but the correct temperature display.
So let's get into the menus of this thing
because it gives a real insight
into how this was going to work
in terms of what you could program
and what you could do with this thing.
So you've got a couple of pre-programmed options here.
You've got single, which says 93 Celsius,
nine bars, 25 seconds.
And so 25 seconds says that they're looking
at like a fixed time, not a fixed volume of brewing.
A few machines work this way,
they'll brew for a fixed amount of time
and you would adjust your grind
to get the flow rate that you want
to get the kind of yield you want in that fixed time.
Interestingly, the double pre-programmed here,
same temperature, same pressure, 35 seconds,
which I hope is not suggesting that a single
is just like a double cut earlier,
because that would be wrong.
You've then got settings, which we can jump into.
And then I think you've got like a more manual thing
where it's going to do the temperature
and pressure that you want,
but it won't cut off after a certain amount of time.
Settings is again, interesting.
So you can adjust your PID values
around a bunch of different stuff.
So if we jump in, obviously your brew temperature,
tap in there, so you can set your proportional gain,
you can set your integral gain,
or you'll set your derivative gain.
I don't know how to program a PID, if I'm totally honest.
So I'm not going to mess with these values.
Now on the back screen,
there's a little bit of immaturity in this machine.
And so the fact that it says back that ass up,
I don't know, I don't think you should build that
into a daily espresso machine.
It just feels kind of weird.
It just feels a little bit childish, if I'm honest.
It's not that I'm against immaturity,
but you know, I don't know.
It's just a weird little thing.
I don't know if it would have made it to production
if it was just for the beta units or something,
but there it is.
We should make some coffee.
Pull, I think we're just going to pull manual.
I think that's a good idea.
I'm going to brew at 94, I think.
And I'm going to brew at eight bars, because why not?
And there it is, programmed very quickly.
So from a programming perspective,
this is kind of a joy to use.
This really is, but can it make good coffee?
So let's brew, and this is interesting.
So right now it's kind of getting prepped.
You can see negative pressure.
It's sort of seemingly filling the hydraulic circuit,
getting ready.
Timer reads zero still.
Once we switch to positive pressure,
okay, timer's on.
Only about four bars of pre-infusion,
and then it's going to really spike it up.
So it's overshot there,
so we're going to need some PID control work done.
Drops back down to about eight bars, which is good.
Temperature a little high.
Now we pulled a little longer than I'd want,
but I think it's still worth tasting.
Now, actually, just on the temperature gauge here,
it's dropped down to like 80 degrees Celsius,
so I wonder if it's got a bit of a declining profile here
with the heating stuff.
Now, some of you, like me,
were you a backer watching this?
Are you thinking, did I miss out, or does it suck?
Let's find out.
It's got a weird acidity to it
that makes me feel like the brew temperature
is not quite right.
Actually, we can probably test that.
We do have a SCASE device
that we can test the brew temperature with
and see what it's actually doing
in terms of temperature and pressure
delivered onto the puck.
Maybe we should do that.
Let's repeat this and actually see what's happening
inside the machine
versus what it's telling us is happening.
Okay, so we've got a SCASE 2 device set up here.
This has a temperature probe inside there
and a flow restrictor to mimic the flow rate of an espresso.
That's running to this laptop.
You can sort of, in a second,
see what this sort of screen recording of this is.
You also have a pressure gauge here,
which is the sort of pressure,
including the loss of pressure through the puck.
So typically, this would read about a bar lower
than you would see at a pump.
So if it was nine bars at the pump,
you'd see eight at the group head
because you're losing about a bar of pressure
through the cake.
Anyway, let's see what happens when we pull a shot.
So we're coming up to pressure here.
Coming way up.
Now that's quite a big discrepancy.
Oh, that's interesting.
So looking at the brew temperature,
it briefly spikes to about 91 degrees Celsius
and then just crashes back down really, really quickly.
And so it's brewing way below sort of target temperature
almost the entire time.
There's a little leakage here, pressure issue here.
I think there's a problem with the group head solenoid.
Touch on that in just a second.
Hang on, let me stop this.
Now this was a beta testing unit that was out in the wild.
So this has been sort of shipped, so to speak,
to somebody to test.
And so you'd expect it to perform a little bit more
within spec than it was.
So what you were seeing was it was thinking
it was delivering about nine bars of pressure.
What you were seeing at the puck though
was substantially below what you'd expect.
The temperature piece though, really,
really just miles away.
And so I don't know at this point
if it just needs its settings tweaked,
but that kind of makes me a little hesitant
to think that's the problem, that's the issue.
Or if maybe it was never ever gonna do
what they really needed it to do.
It's an interesting artifact of its time.
It's definitely got some interesting ideas in there.
It's a sad story that they didn't get to deliver
what they wanted to deliver,
but I'm left wondering if they were ever gonna deliver
something that would have lived up
to the expectations of the community
and to the sort of technical expectations
of really good, accurate espresso brewing.
Now, to caveat myself again and over and over again,
this is sat, you know, not making much coffee
for quite a long time.
I think there's something trapped
inside the group head solenoid.
It's a little sort of electromagnetic valve
that pulls a piston up and down
to control a three-way valve.
And sometimes stuff gets trapped in there
causing a kind of leak of pressure
from the boiler itself sort of through the system.
And you'll see the group head dripping sometimes.
So, you know, this is not maybe in peak condition,
but what its sensors are showing
versus what we're reading with sort of mechanical readings
or, you know, accurate readings of temperature
are quite a long way apart.
And that would be a huge concern for me.
Even if the math was a little better,
it was, you know, staying on target
when it was supposed to do a little bit more accurately,
it just seems to not really know what it's doing
in that particular regard.
And I don't know why that may be the case.
I don't know if that's bad build,
if that's bad componentry, if that's something else.
They're using cheap parts by their own admission.
Maybe that was the issue.
There's lots I like about this.
I like the form of it.
I think it looks kind of cute.
It's boxy in a charming way.
I like the interface.
It's simple.
It let me do what I needed to do,
but I'm not sure it delivered.
You know, I don't want to salvage this.
This is flawed.
It's not working the way it should,
but it's really interesting.
And it will go to one of my Patreon supporters
because Patreon lets me kind of have a budget
to go and acquire these things, test these things,
and make these videos.
And I'm very grateful to them.
So this will be going to one of them in a giveaway soon.
But I'd love to hear your thoughts.
Were you a ZPM backer?
How do you feel now knowing that for your $200,
you may not have got what you wanted?
Love to hear from you, whether you backed it or not,
down in the comments below.
But for now, I'll say thank you so much for watching
and I hope you have a great day.
