Roasting made easy

James Harper: One day in May, I found myself on the outskirts of Oslo, Norway, at the offices of Roest. They manufacture coffee roasters, and they are the sponsor of this episode. And I was with their machine learning engineer, Callum Gilmour, grinding up some coffees.

One coffee was so lightly roasted that their heavy duty grinders were struggling to cut through the beans.

Callum Gilmour: Will it grind? New podcast series coming to you.

James Harper: Callum finishes grinding seven different coffees, and he lays them out across the table. He boils up some water, pours the water into the coffee grinds, lets them brew for a bit and then he and I, along with Veronika Bolduc, Roest’s head of marketing, go around the table and taste them.

Now I wanna tell you about two of those coffees that were literally side by side.

The first one had these notes of stone fruits. It had a little bit of sweetness, but more than anything else, it was quite acidic, and as an experience, it disappeared very quickly in our mouths.

Veronika Bolduc: It is very short.

James Harper: And then we tasted the coffee right next door to it. This coffee was suddenly very juicy.

Veronika Bolduc: Now we have more of a mouthfeel. It’s more juicy, chewy, you can taste it in your tongue everywhere.

James Harper: It was also so much sweeter, really helped balance out that intensity, and for Veronica, she's now tasting tropical fruits.

Veronika Bolduc: It’s like when you have a pineapple but it’s not too acidic, it’s more on the sweet side.

James Harper: A ripe pineapple, what a wonderful thing. So next door to one another were two very distinct coffees: One was very acidic, notes of stone fruit, and next door to it, very juicy, sweet notes of pineapple. And here is the crazy thing: These are both the same coffee, the same raw green beans. The difference between them was how they were roasted, but that difference, I couldn't believe how small it was.

So what was the difference in temperature between these two coffees?

Callum Gilmour: So, number three we dropped at 206 degrees Celsius, and number four we dropped at 210 degrees Celsius. So four degrees, it’s really not much.

James Harper: Four degrees of difference! It's a 2% difference in temperature, almost a rounding error. And yet, those two coffees were so distinctly different.

So here's the thing about roasting: Tiny changes can lead to noticeably different coffees. And that can be a problem because there are so many things you can play around with, tinker with when roasting coffee. As roasting researcher Mark Al-Shemmeri pointed out…

Mark Al-Shemmeri: …the conversation of whether you should adjust airflow, rotation, speed of the drum or temperature, the answer is all of them will affect it, but in different amounts.

James Harper: When we have so many things to adjust on a roaster and all of them can noticeably change how a coffee tastes, which ones should you focus on? I mean, it's overwhelming. So in the first half of this episode, I will give you an answer to that by speaking with one of the world's leading roasting teachers who has done scientific research to answer, “What is the most important thing to focus on when roasting coffee?”

I know I’m over simplifying but this is needed in a world where people have done the opposite for many years.

James Harper: And then in the second half of this episode, we're gonna look at why roasting coffee consistently is so difficult. As we just saw, just four degrees Celsius can lead to a different coffee, which is why I'm going to crack open Roest’s P3000 roaster, the technology in this machine allows anyone, even me, to roast coffee very consistently and fully automatically. I was so impressed with this technology, I think it's gonna change the coffee industry.

I'm James Harper, and this is The Science of Coffee, a spinoff series from my documentary podcast, Filter Stories, and a journey into coffee's hidden microscopic secrets.

James Harper: Okay, so you love coffee, you want to go deeper into it, make a career out of it. You, like many thousands of others, decide: “I'm gonna try and roast coffee.” So you buy some green beans, you buy a two-three kilo, pretty professional roaster and it arrives on your doorstep. You put it all together and then you're faced with a question: “How should I roast my coffee?”

Now, when you open the manufacturer's instructions, they don't really give you an answer. They just say, “Hey, look, here are all the things you can control on this roaster and the things you can play around with.” It's a very long list. I mean, for example, when I was sat down next to Callum in front of a computer screen at Roest’s headquarters, he showed me all the things that are being tracked and adjusted on the P3000.

Callum Gilmour: We’ve got the fan speed, the RPM, the power, drum temperature, bean temperature, air temperature…

James Harper: Okay? You're starting to get a bit overwhelmed. How should I roast my coffee? I mean, you could get an answer by investing 10,000 euros on a roasting course, but surely there's another way to answer such a simple question. You know what I'm gonna do? I'm gonna get my answer by watching YouTube videos. And you very quickly realize you are not gonna get a straight answer. For example, James Hoffman, the coffee YouTube celebrity, doesn't give a straight answer and basically says you gotta do your research. Okay, let me learn this knowledge that's being passed around in the industry. And so you dive into the YouTube rabbit hole and all of a sudden you are drowning in information. Drop the beans in at this temperature, not that temperature. You should use this sensor, not that sensor. Your roast curves gotta look like this, not like that. If you're roasting in summer, do that. Turn the fans up. No, turn the fans down. Maybe you should play around with the drum speed.

And at this point, you're just overwhelmed. Everything you do to those green beans will impact the flavor. Which of these parameters matters the most? Where should I start? And right now I'm gonna help you cut through it all.

Morten Münchow: If you've been on YouTube, there are a hundred parameters where… no, no, no, there’s these three parameters that are the most important, and this is the sequence and the importance.

James Harper: To cut through the confusion, I spoke with one of the world's leading roasting educators.

Morten Münchow: I'm, uh, Morten Münchow, founder of Coffee Mind. I've been in the specialty coffee business for almost two decades.

James Harper: I really admire Morton because he straddles the divide between the scientists and the craft of coffee, you know, the people who roast and brew coffee,

Morten Münchow: I've been teaching more than 1,600 students in small groups of six to eight in London School of Coffee and Coffee Mind Academy in Copenhagen, and also, in training centers globally. So you're at home, you've got your bag of green beans, and your roaster is set up. How should you roast those beans? There are so many things I can adjust and play with. But which thing matters the most? Should you focus on the fan speed, the drum speed, how long the beans are roasting, how quickly the beans get hot, what temperature the beans are at when you dump them out?

James Harper: Now Morton has done some research with the University of Copenhagen to answer that question. He roasted the same green bean in lots of different ways. They found some taste panelists, trained them to identify things like acidity, fruitiness, calibrated them, and then served these coffees blind using a red light. So you couldn't even see the color of the coffee. He then did some very fancy math.

Morten Münchow: You're looking for the vector through the dataset that has the biggest amount of explained variance.

James Harper: And basically figured out what are the things which have the biggest impacts on a coffee's flavour.

Morten Münchow: Where do we have the biggest variance? That's where we need to focus first.

James Harper: And at the end of the process, he figured out what is the strongest predictor for what a coffee is gonna taste like and the strongest predictor is the color of the roasted bean.

Morten Münchow: So in our research overall, we found that color in itself drives 80% of the possible flavour modulation of a green coffee.

James Harper: There it is: 80% of a coffee's flavors are determined by the color of the roasted coffee bean. And I love this because it's just so beautifully simple.

Morten Münchow: People assume that it's complicated, so they assume that it's kind of a hundred parallel processes, and if you do it like this, then this part is enhanced and this is pressed or, or something like this. I don't think it is. So basically it's just coffee getting darker.

James Harper: So when you're there in front of your new roaster with your bag of green beans, the most important thing for you to do is roast a whole bunch of different coffees to different shades of brown, then grind them up, taste them all, and decide on the one you want. And if this sounds too simplistic, well let's just take a step back and reflect on this number, 80%. Morton's research suggests 80% of a coffee's flavour is determined by how dark it's roasted. That 80% difference is the difference between a coffee tasting very vegetal, fruity, and burnt. And that is an enormous difference. Now, it's not the whole story, but it's the thing that makes the biggest difference when roasting coffee. It is the key thing you should focus on first.

Morten Münchow: So first get the basics straight and then build on top, rather than get confused in all the details, right?

James Harper: And one small, but important caveat, is that it does matter how fast you roast your beans. Don't go extreme.

Morten Münchow: Then of course you can burn it by roasting it way too fast. Or you can severely bake it by roasting it way too long, but those are extremes. Stay outside extremes, then you can actually understand it as, colour.

James Harper: Okay, so let's say you've done step one: You roasted the same coffee to many different shades of brown. You tasted them all and you chose the one you wanted. 80% of the work is now done. Now it's time for you to focus on the second most important thing.

Morten Münchow: And once I've found the preference for color, then I could play around with preference for development time.

James Harper: Morton's research suggests that 15% of a coffee's flavour is determined by the development time. This is the moment from when the beans first start cracking, popping like popcorn, and when you dump them out of the roaster. You can listen to the previous episode to learn more about that.

Morten Münchow: That's why development time experiments are the next thing you're doing.

James Harper: So you have your bag of green beans, you have your new roaster. You roast mini coffees to different shades of brown, you taste them all, choose the one you like. That's step one. Step two: Development time. The color you want. Try to get to that colour again, but in different lengths of time. Try to add an extra 30 seconds between the first crack and the color you want. You can also try getting there 30 seconds faster. Then you just have to taste all the samples and choose the one you want. That is the next 15% of the coffee's flavor figured out.

But why is step two, playing around with the development time, only worth 15%, whereas step one, the color of the beans, accounts for 80% of a coffee's flavor?

Well, at Roest’s headquarters, Callum, Veronika and I did another coffee tasting. Two coffees are on the table. They look to me an identical shade of brown, but there is a difference in their development time. One was like a normal roast, but the second is what roasters called a baked coffee.

Callum Gilmour: So, it took more time. The energy input was lower to get to the same degrees. An extra three or so minutes.

James Harper: Now none of us knew which coffee was which, we had anonymized them. And when I tasted them, the difference was pretty small to me. But after enough sips, we think we found the baked coffee because it was a little bit astringent and had this subtle papery note, whereas the other one had more vibrancy to it.

When we did the big reveal, thank God we were right.

Veronika Bolduc: Yay, we can keep our jobs.

James Harper: So yeah, that's the difference 15% makes. It's not dramatic, nowhere near as dramatic compared to how dark you roast your coffee, that first 80%.

Morten Münchow: It does make a difference, but not compared to your decision on when to stop it from a roast degree perspective.

James Harper: So that's development time, step two. And look, once you've done step one and two – you've decided on the color brown you like, and you've played around with the development time – you are now 95% done. This now takes us to step three, that final 5%. And according to Morton, that final 5% of a coffee's flavor is determined by how long it takes you to get to first crack.

Morten Münchow: If people don't add milk, you could also play around with where to put the first crack.

James Harper: And Morton argues, look, if your customers are gonna be adding milk to this coffee, don't even bother with this third step, A 5% difference in flavour is really, really subtle. Adding milk just drowns out those subtleties. No one is gonna be able to taste the difference. Morton argues you should only do that final step, playing around with the timing of first crack, if you're gonna be drinking this coffee, black, you know, a black espresso, a black filter coffee.

Okay, so let's say you've done steps one, two, and three, but you're not happy with the flavour of the coffee, you might think to yourself, “Hmm, did I roast it wrong somewhere? What if I adjusted the fan speed? What if I did this?” No, no, no, no, no, no. For Morton, the answer is very simple.

Morten Münchow: The selection of green is by far the most important decision, and there's only so much you can do with roasting.

James Harper: If you can't get the flavours you want after going through steps one, two, and three, you have to swap out your green beans for something else.

Morten Münchow: You have actually quite limited possibilities with roasting compared to the magnitude of difference you can introduce by selecting different greens.

James Harper: I have tasted an enormous variety of different coffee flavors over the years. Mango. Jasmine. Dark chocolate. But if I want those flavors, I can't always find them through the way I roast the coffee. I need to buy the green beans which have that flavour potential in the first place.

There it is, that's Morton's framework.

Morten Münchow: So it’s 80, 15, 5.

James Harper: I really love Morton's way of thinking here. You know, myself as a storyteller, documentary maker, I collect dozens of hours of recordings, interviews, and then I spend like a hundred hours going through it all, editing, editing, editing to give you a story, which I hope is simple, interesting, and you walk away having learned the science of roasting.

What would be so much easier for me is if I just gave you those 20 hours of raw recordings and said, “Hey, in there somewhere is the science of roasting. Good luck.” That's kind of what it's like trying to learn how to roast on YouTube. You're just flooded with information and you don't know what really matters. What Morton has done here is what I do with this podcast. He's got to the essence of the story, created a framework that simplifies our very complicated world. Follow these steps – 1, 2, 3 – and anyone who has never roasted coffee before can get pretty decent cups.

However, unfortunately, Morton's framework doesn't help you solve one of the biggest problems in coffee roasting, because here's the thing: You can know the shade of brown you want, the development time you want, and how long you want to get to first crack, steps 1, 2 and 3 of Morton's framework. Good luck though, nailing that every single time.

And this is what I wanna explore next in the episode. Roasting coffee consistently. If you wanna explore Morton's framework further, I've linked his paper in the show notes and his website.

Let me tell you a little story about consistency. I moved to Berlin in 2016 and I worked as the wholesale manager for a very well-known roastery. Now, September had rolled around and it was Berlin Coffee Week, and the owner wanted to put on a brewing competition, and the winner would get a very fancy hand grinder. For this competition, the owner chose what was the best coffee we had in-house, a Kenyan peaberry. Leading up to this competition, these beans tasted unbelievable. It was like marmalade, ah, I can still remember it now, it was so good. So competition day rolls around. I myself also fancy competing because who doesn't want a fancy hang grinder? So, I'm behind the bar in a row with all the other competitors. We're boiling our water, grinding our beans, measuring everything out, making our filter coffee so delicately. I put my finished coffee in a cup for the judges and there's a little bit leftover, and I take a sip and, oh, oh, oh, dear, this does not taste like marmalade at all. This tasted like tomato; I was getting these vegetal notes. This coffee wasn't as sweet as I remembered it. And I thought to myself, “Damn, I've brewed this badly.”

But when I look to my left and right, I see all the other competitors tasting their own coffees with frowns on their faces. No, something was not right with these beans. These beans needed a bit more development, you know, maybe another 20 seconds in the roaster before being dumped out. Now, the owner was very embarrassed, you know, the best and brightest of the Berlin coffee scene had all come along. And all walked away being like, “Hmm, this is not a well roasted peaberry.”

But the thing is, I knew the roaster knew how to roast these beans. Well, previous batches were so good, and the reason this was an inconsistently roasted coffee is because the roasting itself was done very manually. You know, the roaster would be stood there all day, adjusting a whole bunch of different parameters, adding more flame, pulling back the flame, sometimes the door of the roastery would swing open and in winter a big gust of cold wind would rush in and that would affect the roast and the poor roaster themselves. They had been standing there roasting coffees for hours on end since the crack of dawn, and honestly, by the time they got to midday and they're roasting this coffee, they were tired, they were bleary-eyed. It's like they weren't even looking at me. They're looking past me into the distance. and this problem is not exclusive to that roastery I worked at in Berlin. I remember the same thing happening in Australia. Melbourne, 2015, I went to a public coffee tasting at one of the most acclaimed coffee roasters in town, and we tasted this experimental coffee that blew my mind, I was so desperate to taste it again. I went back the next day and I ordered it this time as a customer at the bar and what I got, it was not the same coffee. The vibrancy wasn't there. It was like a poorer cousin of the coffee I tasted the day before. Roasting coffee consistently, it's hard. For the longest time, it has been one of the big headaches for coffee roasters. But these stories took place almost a decade ago. Things have changed a lot recently.

I was really amazed as Callum showed me how to operate Roest's P3000. They claim this roaster can roast coffee more consistently than us humans and entirely automatically. So I want to tell you the story of how this machine came about and then we're gonna open it up and look at the technology that enables it to roast coffee consistently and fully automatically.

So, this machine is the brainchild of Sverre Simonson.

Sverre Simonson: So I'm, um, Sverre and I started this 10 years ago with my brother Trond.

James Harper: So after high school, Sverre had to do his compulsory military service. He hops on a bus and he is driven to the frozen north of Norway, up near the border of Russia.

Sverre Simonson: We were moving people across small lakes or rivers.

James Harper: You're like the engineering team. This is what the engineers do in the military; they help the military get around.

Sverre Simonson: Yeah, pretty much.

James Harper: And dressed in his military fatigues, overlooking the frozen lakes in the arctic circle, Sverre falls in love with coffee.

Sverre Simonson: Although the coffee was absolutely horrible, I sometimes mixed it with the hot chocolate in the field regions. There was still something nice about a hot cup of coffee in the morning.

James Harper: And his love of coffee follows him when he decides to do a master's degree in mechanical engineering after his military service.

Sverre Simonson: So during my study, my interest for coffee grew, and I continued upgrading my coffee gear, but I never really found a coffee that I loved, so I decided to build my own coffee roaster, which is obviously the next step.

James Harper: He focuses his master's thesis around building his coffee roaster and as part of the project, he went to speak to some of Norway's most well-known coffee roasters.

Sverre Simonson: Timbo and Morten in Nordic Approach.

James Harper: And they tell him, “We are having issues with our sample roasters.”

Sverre Simonson: And that's where the idea of making a sample roast came about.

James Harper: Now a sample roaster is a small coffee roaster that cooperatives, roasters, exporters, importers use to get a gauge on what an entire container of coffee tastes like by just roasting a little bit of it. And these experienced specialty coffee professionals told Sverre they had two big issues.

Sverre Simonson: So one of the first things they mentioned was obviously they had issues with consistency.

James Harper: They were struggling to roast their coffees exactly the same way every time, and also they were annoyed that these machines weren't automated.

Sverre Simonson: They wanted something that they would spend less time operating.

James Harper: And while Sverre is working on this machine, his brother Trond is like, “Hey, we should totally bring this thing to market. Let's team up, you and I, brother and brother.”

Sverre Simonson: Trond arranged the funding and I built a workshop in my basement. And then I became an Airbnb superhost in order to get some extra income in order to pay the bills.

James Harper: So they developed and then started selling this sample roaster and they proved they could roast coffee automatically and very consistently. But it's a sample roaster – it can only roast small amounts of coffee at a time. They decided to scale it up to the P3000.

Sverre Simonson: So the P3000 is a full electric roaster, obviously. It has a three kilo capacity, but it's designed as a 500 kilo per week capacity roaster.

James Harper: Sverre tells me the P3000 will be popular with, for example, the coffee roastery that I used to work for. If they had the P3000, they wouldn't have messed up that batch of Kenyan coffee for the competition.

Sverre Simonson: If you have a specialty coffee roastery, uh, maybe you have a 30 kilo main roaster, and then once in a while you get some really good coffee and it's gonna cost you a lot to mess it up.

James Harper: Let's now open up the P3000 so I can show you the remarkable technology that allows it to roast coffee fully automatically and very consistently.

Okay, so let's imagine you wanna roast some coffee. You dump the green beans into the roaster. Now to roast coffee consistently, batch after batch, after batch, it's really important you take out your beans at the exact same moment in every roast. As we saw in the last episode, I showed you a roast that took about seven minutes and for the first five minutes, not much happened, the beans just got hotter. And then we had our first crack – the beans exploded like popcorn. And for many roasters, this is the key moment. They start the timer at first and to be consistent, they try to stop each batch of coffee at, for example, 30 seconds after first crack. But here's the problem. When is the first crack, first crack?

So here's an example. I recorded this audio just next to the machine. This is what you would be hearing if you were roasting this coffee yourself and trying to determine when the first crack begins. So you're the roaster, when is the first crack?

So, did you hear the first crack? It's hard to hear. It's really, really subtle, you have to be there like a hawk listening for every single pop. Now, imagine if you're the head roaster, you've been roasting batch after batch for hours. It's a busy production floor, people are running around asking questions of you, and then you have to really sit down and focus on these very subtle sounds. And here's the really crazy thing: You could absolutely have said that the first crack began in earnest at the end of that audio. Guess what? That audio snippet was 20 seconds long. That is long enough for a coffee to taste distinctly differently from batch to batch. Sverre, in order for his roasters to roast consistently, needed a way to define precisely when the first crack begins.

Sverre Simonson: If you have a different definition of when it starts, you will also have a roast that is a little bit darker or a little bit lighter depending on who's using the machine.

James Harper: Callum explains to me they found the solution first by sticking a microphone into this flaming hot oven, and then they process the audio so the cracks themselves would be particularly audible, and then they did a whole bunch of very wonky computation.

Sverre Simonson: What they are is the envelope of the time-power spectrum of a signal.

James Harper: If it's this loud for this period of time, for these specific sound frequencies, we are defining this as the first crack. And so, regardless of who uses the machine and how tired they are, the first crack is consistently defined. But of course, to roast consistently, it's more than just knowing when the first crack is. You need to know how hot the beans are getting and at what temperature they need to be dumped out. Now for the longest time, roasting technology has been pretty simple, especially when it comes to temperature readings. Roasters would have a temperature probe called a thermocouple, which is basically an electric thermometer that would sit inside the roasting drum and it would measure the temperature of the beans as beans hit the probe. Now, the problem with thermocouples, these electronic temperature probes, is that they only get a reading when beans literally hit them, which is not all the time and when they do, there's a relatively high margin of error in the reading plus there's a lot of noise in the reading. The probe could be measuring the temperature of the air rather than the beans.

Callum Gilmour: On this machine we’ve actually got three temperature-sensing technologies.

James Harper: When I visited Roest’s R&D lab, Callum was experimenting with not one, not two, but three different temperature systems.

Callum Gilmour: Having this kind of grid of temperature-sensors with various technologies means we can use the benefits of all of them and get an average that’s an accurate representation of the temperature that’s there.

James Harper: More types of temperature reading tells the software more accurately, “This is how hot we think the beans actually are.” This helps the software track the roast more accurately, control how much heat it's applying, and again, this reduces inconsistencies between batches.

Callum Gilmour: We do have a window at the back here where we have a camera unit.

James Harper: And what I found really surprising is that Roest is developing a technology I have never seen in a roaster: a camera system.

Callum Gilmour: We are gonna use the camera to measure the surface bean color.

James Harper: Now, for the longest time roasters have tracked the color of the coffee beans, but the way they would do it is kind of cumbersome. They would use this scoop, technically it's called a trier, which they stick into the roaster, pull out a couple of beans, look at the color, judge whether they're on track or not, and then adjust the roast. Again, very manual, and very prone to human error. What if you're roasting at the crack of dawn, compared to midday? The amount of ambient light could change the color of the beans. The roaster could misinterpret how dark the beans have actually got. And so by developing a camera system, Roest’s software can accurately track how dark the beans are getting over time, which then helps produce more consistent roasts. And I was really impressed with the technology they developed for this.

For one, you need a camera that can shoot incredibly fast because those beans are hurtling around that roasting drum so quickly.

Callum Gilmour: We’ve minimized our energy difference, brightness difference between frames to under 1%, which we’re happy that we can take good readings from.

James Harper: Their camera system is so fast that they actually had to develop their own LED lighting system.

And now here's the final piece of technology I wanna tell you about in the Roest P3000. I spoke with the specialty coffee roasting consultant Scott Rao, and he told me about a problem he'd had for many, many years with smaller roasters.

Scott Rao: The problem is you can use the same inputs every day and get a slightly different output every day, because you might have changing ambient conditions, you might have changing outdoor conditions.

James Harper: Essentially, roasters are made of metal, and the heat of that metal changes throughout the day, which can lead to inconsistencies. As we saw at the very beginning, the same coffee that was roasted just four degrees hotter, was distinctly different from the coffee before it. Scott Rao developed a system called the Between Batch Protocol, which is basically just every time you roast a coffee, the roasting machine itself, the metal, it's always at the same temperature.

Scott Rao: We're trying to come up with a system to reset that metal's energy and most machines until recently didn't really help you do that.

James Harper: And when Scott was presenting this at a conference over in America, the Roest team was in attendance, heard it, and then created a Between Batch Protocol for their own machines.

Scott Rao: In the past year, Roest came to understand the importance of the warmup in Between Batch Protocol.

James Harper: So in between roasts, the machine will heat itself up, cool itself down a little bit, heat itself up, and when it's ready for the next batch of coffee, the machine itself is always at the same temperature.

All of this is some of the remarkable technology I saw for myself when I visited Roest. And they were very keen to explore any way they could make their roasters more reliable, more consistent. So if you have any ideas, get in touch with them: link in the show notes. I'm sure they'll be very happy to hear from you. And if your idea involves machine learning and prediction models, don't worry – they're already on it.

Sverre Simonson: And the goal with P3000 is that once we have launched a machine and we start getting data from our users, we can see what type of profile works well on what type of coffee. And the more data we get, the better we can give a customer an idea of where you should take this coffee when it comes to profiles. You're basically going from not knowing how to start roasting this coffee to getting a recipe from the machine that will get you almost to the end.

James Harper: So, after visiting Roest, I came back to Berlin with my head spinning because this kind of technology just changes the game in the coffee industry. The technical barriers to roasting are gone, you don't need to invest as much money on learning how to roast coffee, because these machines roast automatically, you could roast coffee part-time while holding down a full-time job. Becoming a coffee roaster has never been easier or cheaper, and the end result of all of this is competition for roasted coffee is going to get a lot hotter.

And here are three changes I think we're gonna be seeing sooner rather than later. So, let's talk specialty coffee shops. The owners of specialty coffee shops get into it mostly for passion. People who get into specialty coffee to make a lot of money quickly realize the error of their ways. This is a passion business. And the owners of these coffee shops would buy roasted coffee from established roasting brands because they didn't have the time and the money to roast for themselves. But now they do, so I think we're gonna see many more specialty cafes roasting coffee for themselves.

I also think we're gonna be seeing the same roasting brand operating across many more geographies. Right now, if you are a London-based roaster and you wanna come and, you know, set up a business in Berlin, It's important for you to offer the same coffee here in Berlin as you offer it in London. Now, one way you could do that is to roast all your coffee in London and ship it to Berlin. But good luck trying to get roasted coffee reliably, cheaply into Berlin in a post-Brexit world. But now roasting is so easy. you could put a fully automatic roaster in Berlin Roast the exact same coffee you're roasting over in London. When you buy your green beans, half of them get shipped to London, half of them get shipped to Berlin, and they get roasted in their respective cities.

So that's why I think we're gonna be seeing roasting brands operating across many geographies. You might see a roaster that's based in Sydney and New York roasting, the exact same coffee in both places.

To push that further, I think this now opens up a really cool opportunity for coffee producers, the growers, to sell their coffee more directly in the global north. There are many barriers trying to sell coffee roasted outside of Europe into Europe. But now it's much easier. Let's say you run quite a large farm in Colombia, you have tourist groups coming in all the time. Let's say you get a visitor from Germany. You really click, you really bond with them and they get excited to sell your coffee in Germany. Well, now that German doesn't need to have 10 years of roasting experience; they can roast your Colombian coffee easily, reliably, and sell it locally. and so the big shift is that.

To have a successful roasting brand in the past, you needed to tick the boxes of, you know, just good business. You had to have good sales channels, marketing, customer service, et cetera, et cetera, as well as maybe a hundred thousand euros of investment and a lot of technical roasting knowhow. But now the financial barriers, the technical barriers, they're so low, they're not a barrier at all anymore. The roasting businesses that do well in the future will be only down to the classics of business, you know, sales channels, marketing, customer service. So those are my thoughts. What are yours? Where is the coffee market going? If you have ideas, I would love to hear them. You can reach me on Instagram at Filter Stories podcast or search for me on LinkedIn. Link in the show notes.

And if you are interested in the P3000, check it out, I have a link in the show notes. Get in touch with the Roest team and tell them you heard about it on this podcast. And to be clear, I don't make a commission on any sales. In the next episode, we are going to confront the end of the world. Civil wars, refugee crises, a scorching hot climate, and maybe a solution to it all – farming coffee organically.