One evening, after admitting to the team I had read an article about using laser diffraction to measure the particle size distribution of coffee grinds (it’s a Patreon post by astrophysicist Jonathan Gagne who also made an app to measure grind size from a photo), Greg asked a question:
At Project Vault we’ve gone ‘Office First’, and I’m spending more time in the office. Working in the office with real three dimensional human beings is great but I’ve had two real problems:
So after Greg posed the question my immediate thought was: does this mean we’re going to have good coffee in the office?!?
By way of background: I’m quite into coffee. What I lack in a sophisticated palate I make up for in enthusiasm. I genuinely believe throwing myself at a few coffee-related projects helped drag me through the first lockdown. I built myself a coffee station above the kitchen bins, and figured out how to measure pressure, draw graphs and fire webhooks using arduino so I could add instrumentation to my classic lever espresso machine.
This is serious business and will require some structure:
In the spirit of ‘Always Be Shipping’ - here is Part 1. Parts 2 and 3 to follow.
'If I had only one hour to save the world, I would spend fifty-five minutes defining the problem, and only five minutes finding the solution.'
Easy, right? Drinks you make from ground coffee beans. We’re after the best one of those. Instant coffee is off limits. (I’m not an Instant hater, I just define it as a different drink.)
There're a few distinct ways of making coffee which result in quite different drinks which are hard to objectively compare against each other.
The most common are:
E.g. French Press / Clever Dripper / Aeropress
Mix ground coffee in hot water. Let it sit to brew. Filter out the grounds when it’s brewed for long enough so you’re not chewing on ground coffee
E.g. Drip coffee / V60
Typically put ground coffee in a conical filter. Slowly pour hot water over the coffee so it percolates through and into a receptacle underneath.
The thing you get out of an espresso machine and what you use as a base for flat whites/cappuccinos etc. Force hot water at high pressure through a compressed puck of ground coffee
Each of these methods makes quite a different drink. Espresso is a much more concentrated drink than either of the brewing methods due to the high pressure involved. Brewed coffee is usually lighter and can be more subtle and fruity.
Geeky aside: The difference between Immersion brewing and filter brewing is more subtle. Gagne has a good article explaining why they’re different -- it’s due to how the concentration gradient changes over time.
Now you have three methods, each of which can make amazing but different drinks. How are we going to make the BEST coffee?
I’m going to argue that to say we make the best coffee it needs to be espresso. Why?
(Btw, I’m not saying espresso is the best coffee. That is a subtly different statement.)
Geeky aside. For the purposes of this I’m somewhat arbitrarily setting the rough bounding parameters of espresso being ‘a drink made from ground medium-light to medium-dark roasted beans, extracted with a peak pressure >6 bar and with a ratio of ground coffee to water of between 1:1.5 to 1:4 by weight.’
There’s two tricky bits to this: taste and distribution.
One of the limiting factors when people get really geeky with coffee is that the ultimate goal is making tasty coffee, and taste is very hard to monitor objectively. There is also no agreement of what the best espresso should even taste like, and different styles go in and out of fashion. And lets not forget that:
You can see how this could go wrong.
We’re going to make lots of cups of coffee, all of which will be different. When we say we have the best coffee, are we referring to the quality of the best cup of coffee we make? Or the worst? Or the average?
The quality of the best cup is not a great measure. With poorly controlled processes and lots of randomness, if you make enough coffee occasionally the stars will align and you’ll make an incredible cup. (Within reason. If you have mistakenly ground soil instead of coffee beans you will never make fantastic coffee.)
You could make the quality of the worst cup the benchmark. If you look at the quality of the worst cup of coffee we make, is it better than the worst cup of coffee from any other startup? Doing this would prioritise error-proofing over all else. Even one bad cup would ruin everything. We’d have to compromise a lot to ensure it was impossible to mess up the process. On this metric a Nespresso would handily beat the coffee I make at home. I’ve made some truly awful coffee.
That leaves us with something in the middle. I’d probably go with the quality of the median cup, and if I could add a second metric would look at how tight the Inter Quartile Range is.
In the world of espresso it’s the consistency that’s hard. Espresso is typically made using water at 9 bar of pressure -- roughly the equivalent of the full weight of a VW Golf pressing down on one of your hands. Ramming 9 bar of pressurised, almost boiling water at a small disc of ground coffee is a highly effective way of highlighting any small differences between how you’ve prepared two seemingly identical shots.
By choosing the median we care about not just the best coffee but the quality of the average coffee. We can’t just make a few amazing coffees and the rest be rubbish, we have to be consistently making great coffee.
Who cares? Will we even be able to measure the quality of the median coffee? Even if we can’t measure the median, knowing we’re optimising for the median coffee will drive the selection process and equipment.
P.S. Once we’re up and running, if anyone wants to challenge our assertion of having the best coffee then let me know. I’ve got some initial thoughts on how we could try and semi-objectively compare the quality of our respective median coffees.
Ok I’m tired of definitions now. Lets just say any company 5 - 500 people whose business purpose has no connection with coffee (e.g. coffee shops, roasteries, etc. are not included!) and where not everyone making coffee is a coffee geek. I was going to exclude companies which have professional baristas but I think maybe I’ll leave those in. We need ambition.
Coming soon. Part 2 and 3: Now we’ve defined best, how might one achieve this and does it make any sense for us to pursue this audacious goal?