The Home Roaster Business Case

I like good coffee.

I like it so much, that I don’t mind spending $26 NZD on a 250g bag. This here is from one of my favourite roasters and it’s absolutely delicious. It’s a fruit bomb of a coffee.

The kg price of this bag sums to $104.

Now here is something interesting I found out. The green bean supplier below sells the exact same coffee, un-roasted, for ~$34/kg.

That’s a whopping $70 dollar difference. Roasting these beans will lose about 12% in weight, but the difference is still $65 per kg.

At this point I’m obviously asking myself:

1. How hard can roasting coffee be?

2. How much is a coffee roaster?

I’ll get to #1 later, but I found a second hand home roaster for $900.

Say I’m saving $65 per kg.

900 ÷ 65 ≈ 14 kg of roasted coffee

I would pay for that roaster after 14kg of roasted coffee!

That’s less than my annual consumption. We drink about 4 cups a day, 15g each making this 1,8kg/month i.e. we’d after ~8 months.

The model I got is from an NZ company called Kaffelogic.

I went for this one mainly because it was second hand, it’s super small and it’s an air roaster, which seem easier to handle.

How Hard Can It Be?

This of course is my gamble. If it’s too hard, i just wasted 1000 dollars.

I have zero experience in roasting, I just enjoy coffee, grind my beans and do filter coffee.

I’m pretty sure it can’t be that easy. If this is where all the money is, then there has to be some justification. Maybe it’s the expensive machines if you want to produce at scale.

But then, my business case above shows that pretty much anyone who buys specialty coffee, would be able to pay for a home roasting machine in no time.

Kaffelogic has been around since at least 2018, producing home roasters. So how come I’ve never heard of this?

It get’s even better. The Kaffelogic Nano comes with a whole set of profiles that are seemingly plug and play.

The guy who sold me the roaster packed some green beans.

The key points here are the process “Natural” i.e., how the raw beans have been processed and altitude. Beans grown at 1800 MASL are much denser and that needs to be taken into account on the roaster.

Accordingly, Kaffelogic comes with something like this.

So to roast above coffee for filter brewing, I would pick “Filter - Natural - 2000m-2700m”. This then presets to which temperature the beans are heated, when to stop heating, how quickly and likely a ton of other stuff that I can’t yet comprehend.

That’s fine, I’ll just throw in some beans and try the default.

My First Roast

I have no idea how old the beans from the seller are. Green beans supposedly last quite long, but my expectations where quite low.

All I did was take the green beans, weigh 120g and place them in the machine. I then load the suitable default profile and select a roast level i.e., how dark you want the beans to be roasted.

Press play and it goes like this:

I did the same batch again with a darker roast level just to see the difference.

A bit annoying in this workflow that you can’t immediately try it. Roasted coffee needs to rest to degass and release some of the CO2 that has been stored in the beans during roasting.

Expecting nothing, I brew a cup. For the non coffee people, you usually start to smell the flavours when you start pouring water over the grounds and sure enough I smell fruit. FRUIT! On my very first roast. Fruit is typical in these Ethiopian beans, so that’s great.

In the cup it even tastes like fruit which kind of blows my mind. It’s not perfect and has a little bit of a grassy note to it, which tells me it’s likely not roasted enough.

Either way, I’ blown away. I have roasted coffee and it tastes good.

Not quite as good as Rocket Coffee, but that was my very first try.

The iteration here is slow as now I need to tweak parameters and wait another five days (minimum) for the beans to rest.

The big question. Can anyone do this and will the results be consistent?

Systems

I love systems and I already had this nice system of scanning the internet for the cheapest beans and an AI assisted process of iterating on my coffee brewing, perfecting the process of grinding to the right size and brewing the right recipe.

Now with roasting it becomes this interconnected loop.

1. You buy green beans,

2. you roast,

3. you rest,

4. you grind,

5. you brew,

6. you taste and

7. you repeat.

There is a lot of data involved in that process and it can be used to iterate faster and get us to the perfect roast or perfect cup faster.

The beans will have an altitude, origin, variety and a processing method.

The roaster will have a matching profile. It will also log what actually happens throughout a roast.

It will be configurable via a roast level to determine how light or dark you want your beans to be and which flavours you want to produce in the coffee.

The roasted beans will have a color that can be read via a photo on a white background and classified (it’s called Agtron - a roast-color score for coffee beans).

Knowing the Agtron score of a roast helps us tweaking the parameters next time we roast.

Combine this technical roast information with how you’ve brewed the beans and how they’ve tasted and ideally even someone like me, with zero experience in roasting coffee, can quite quickly get to a decent roast.

If I can do this and it saves me money, the big question is, why doesn’t everyone roast their on coffee.

I get that this might be a specific market and I don’t think average joe who whips up some instant coffee will do this, but I can totally see anyone that owns a coffee grinder with hundreds of dollars to also consider roasting if it’s easy enough.