In my previous post I tried to define what ‘the best coffee of any startup in the world’ might actually mean. In this one I’m going to start to answer how we could achieve it.
I love using things for things other than their intended purpose with occasional unintended consequences. Innovation often happens at the intersections between disciplines so I’m hoping that it might be useful to reframe making coffee as just another process to be optimised .
In process engineering there is a root-cause identification framework called the ‘Five Ms’. used to identify the underlying source of variations in a process.
By systematically working through each of these, you should be able to capture all the potential causes of failure or variation in a process. This is a useful generic framework but rather conveniently the Italians figured this out a long time ago and have their own spin on the ‘M’s specifically for making espresso:
Traditionally you do this kind of root-cause analysis in an Ishikawa or Fishbone diagram. On the right hand side is the problem you’re investigating -- in this case ‘Bad Espresso’ -- and then on the left you list the potential causes. You start with the highest level causes - the ‘M’s, and then break down the sub-causes in ever more granular detail. Doing this properly you should get to sub-sub-sub-sub causes as per the ‘Five Whys’.
This is my quick go at doing the first level of ‘Why’s on the ‘Four Ms’:
That’s a lot of things that can go wrong!
All of these elements of the ’Four Ms’ need to be of the right ‘Quality’ to create great coffee. You then need to be able to repeatedly deliver each element over and over again: ‘Consistency’. There isn’t a single ‘right answer’; there are many recipes - or combinations of all the elements that can deliver great, but different, coffee. In this article I’m going to focus on two of the ‘Ms’: (1) the hand of the operator and (2) the grinder. These, taken together, are more important than how shiny an espresso machine you buy.
If you’re interested in the distinction between ‘Quality’ and ‘Consistency’ and why inconsistency almost was the death of my espresso habit, I’ve expanded on it at the bottom of this article.
I have proved repeatedly that you can have great beans and a great machine yet make absolutely terrible coffee. Most of the battle is won and lost before you touch an espresso machine; it is getting an evenly compressed puck of coffee that’s the hard bit so this is the present focus.
If you watch a great barista it looks easy. The effortless tap of a portafilter (that’s the name of the thing with a handle you put ground coffee into) to level the coffee bed followed by a perfectly flat and level tamp in a single smooth motion.
When you try it yourself you realise it’s not as easy as it looks - it’s a well honed artisanal skill. And this is part of why there is so much scepticism about whether our goal is attainable. The question is:
Rather than relying on artisanal skill for consistency, can we instead rely on process and technology?
At home each time I make a coffee I weigh out the beans I need and put them through the grinder (called single dosing). At Caravan, an excellent craft coffee house near our offices, they have a different method: today I saw the barista grind roughly the right amount of coffee, weigh the portafilter full of ground coffee and then use a tea spoon to remove a small amount to get to the right weight. Either way, using the right amount of coffee (+/- 0.1g) makes a big difference.
Both of those methods are too slow, even for me, if I need to get some work done. Fortunately technology can solve this one. There are now commercial grade grinders which have integrated scales; you program in your desired weight and they have the intelligence to switch off the motor at the right moment so you end up with your target weight of coffee. Unfortunately there is currently a worldwide grinder shortage due to chip production issues which means actually buying one of these is not trivial.
This grinder should make getting a nice mound of fluffy coffee grinds of the right weight as simple as putting the portafilter into the holder on the grinder. It does the rest.
Before you squash the coffee with a tamper, you want to make sure it’s nice and flat. A skilled barista can do this with a few well judged taps. At home you might find me laboriously stirring the coffee with fine needles - there’s no way we can expect the office to do that!
So how can we do it? Honestly - this is the bit I think we might need to experiment with the most. How can we get consistent distribution with minimal effort? I’m hoping the grinder we’ve selected will minimise the amount of redistribution needed, and we’re going to initially try a simple wedge distributor. You put it wedge side down in the portafilter and spin it round. The theory is it evens and flattens it all out. We’ll give it a shot. (sorry)
Tamping is the process of compacting the bed of coffee so it doesn’t fall to pieces when you fire high pressure hot water at it. The tamped ‘puck’ of coffee needs to provide even resistance to the water as water follows the path of least resistance. If there are parts that are easier to flow through, more water will flow through those areas and they will become over-extracted meaning it's more likely we'll extract non-tasty compounds and other areas under-extracted with the tasty compounds still locked inside.
The really key thing here is that the tamp is horizontal - if not more water will flow through the less compressed side.
We can solve this with technology again - but a simpler variety this time. We’re going to try a ‘level tamper’ which has a simple mechanism that makes it practically impossible not to have a level tamp.
After this we should have a nice evenly compressed puck of coffee ready to be locked into an espresso machine so we can force hot water through it at high pressure. Rather than relying on a barista’s skill and muscle memory to ‘get it right’ each and every time, we’re using process and technology to minimise the variation. After all this effort it better taste great.
So will this deliver the best coffee of any startup in the world? Well that’s also going to depend on the other two ‘M’s which are coming up in another post soon…
Within roast coffee beans are chemical compounds. When you pass hot water through ground coffee you can extract these compounds from the coffee and they add flavour and texture to water. In general higher temperature, higher pressure, finer coffee grinds and darker roasted beans all lead to faster extraction of many compounds.
Some of these compounds give desirable flavours and some not, and the game we’re trying to play is to use all the variables we control to try and extract lots of the tasty compounds and none of the bitter, acrid, or otherwise non-tasty ones. We’re also want the same extraction from each ground of coffee; for example if some of the grounds are smaller, compounds will be extracted much faster, and we may start to get undesirable flavours.
(Like everything in this article this is all a (hopefully useful) simplification. Nothing is quite what it seems in espresso; finer grinding doesn’t always mean higher extraction and actually having all your particles the same size can add a whole load of additional complexity)
‘Consistency’ means you can eventually get to ‘Quality’. Without consistency you have nothing.
As an example of the difference between ‘Quality’ and ‘Consistency’; if you are brewing your coffee with water at 98 Celsius you aren’t going to make great coffee as it’s too hot and will taste burnt. No matter how consistent your temperature is - it’s consistently bad because the ‘Quality’ of water temperature is wrong. Conversely if you’ve got a great recipe at 93 Celsius, but your espresso machine’s temperature actually swings between 80 - 100, then sometimes your coffee will be great, sometimes terrible, due to the lack of ‘Consistency’. You can see however that figuring out what the right temperature is, and for that matter many of the other variables, is all a bit pointless if temperature is swinging all over the place.
This is a real problem in some machines and people try to work around it by temperature surfing. If you understand how your machine’s temperature oscillates, you can time when you pull a shot to hit the desired temperature. This is much less fun than actual surfing.
When I started trying to make espresso I had a lack of consistency and it made trying to make great (or even passable) espresso deeply frustrating. Each time I tried the result was somewhat random, and when I changed one thing it was almost impossible to figure out whether the resulting change in coffee was due to what I changed or simply from all the randomness in the system.
When I first started I had:
Given that I didn't really know what I was doing I just got really annoyed a lot. I gave up, put the machine in a box and went back to filter coffee. It was only several years later when I had a better grinder I came back to it; even just removing one source of variation made a huge difference and eventually I went full geek on it and added lots of instrumentation to it so I could really get consistency and make what I deem to be acceptable expresso.
No-one should be dealing with the cost-of-living crisis all alone. We’re building a new service to liberate households from drudgery and make people’s lives simpler and fairer.