This article was originally written for the wonderful people at Coffee Mag in South Africa . This version is slightly edited.
Automation in coffee making is a hot topic. I was recently at the World Barista Championship in Seoul, Korea, an event where the global coffee community assembles, and automation was a point of much discussion. There are opposing opinions. Some certain coffee prophets foresee the end of the barista. In fact, this is often much more than a prediction, it has become a manifesto, one that argues that automation is the march of progress, and that it holds only benefits for the role of the barista, for coffee businesses and for coffee drinkers, and that you would be silly to deny it.
This article is my attempt to do the impossible – to predict the future. The most popular predictions tend to be the ones with the most dramatic clear-cut vision. They often are controversial and challenging. Predictions are most often about engaging us, setting our imagination alight, or pushing an agenda of some kind.
The reality tends to be a bit of let down, or at least a less radical scenario with a more subtle, dynamic and complex truth, one that’s not so far from where we are now.
For me there are two questions:
The first is purely about coffee making theory. Can a machine make coffee as good as or better than the best barista?
The second is less clear cut but arguably more important in predicting the future role of automation. This question is, if the answer to question one is ‘yes,’ how valuable is this for coffee, for baristas, for customers, and for the whole community?
So let’s explore the first question. Can machines make speciality coffee?
There is often a miscorrelation between the increased use of tech in coffee prep and automation. Yes, we have introduced new bits of kit into the progressive coffee shops around the world: we weigh, we are using different, less manual distribution techniques, we are not brewing shots by eye, but if anything, we have added just as much technique as we have taken away. A barista may now have to weigh a dose pre and post grind, distribute it, possibly with a tool, tamp it and then manage the pour through predetermined pressure settings and oversee an ‘out’ weight either manually or by utilizing a built in scale. This process is anything but automatic.
And the beauty of these processes is that they are nimble. If we decide that maybe we want to grind our coffee frozen, like multiple high placing competitors at WBC did based on new research, we can implement this step easily enough. It may not be commercially practical, but the experimental coffee chef in all baristas is always part of the equation.
An automated machine will be a machine that makes coffee the way we decide. Speciality coffee and its members all have varying ideas about what this should look like, and an automated system will to some degree require signing up to one fixed vision.
I think speciality coffee holds flexibility very dear, which makes sense in multiple ways. The ability to trial a new grinder next to my machine (or a dosing tool etc.) is something I wouldn’t give up unless the benefits were huge. I think this is a well-shared feeling, and there are multiple other reasons that I, as an operator, would be concerned about the change.
Anyone who has been involved in the development of a physical, functioning product, (and most who haven’t) will understand that these products cannot be easily updated like an app. There is a multiple year process of R&D and supply chains that need to be managed. Altering a major element of the process means going back to the drawing board. The company making the product also has to see commercial validity in doing so. Will there be enough customers who want this product, who value these details and who want to pay for it?
So, in answer to question one, it’s pretty clear that the super-auto machines are fast approaching the ability to make coffee indiscernible from that of a good barista. Acknowledging a certain expenditure and inflexibility of brewing, the machines will keep on rolling, and will at some point be able to make coffee as well as (and probably more consistently than) a good barista. Not the singular best coffee that it is possible to make, but definitely serviceable for speciality coffee purposes. But what does this mean for the barista?
This is where we move to question two. What is the value proposition?
If I have a barista called Jen who is a highly engaged, experienced barista who finds coffee making rewarding and likes to share the process with curious customers and fellow colleagues. It is an involved craft that she finds rewarding as a physical process.
Then I have Frank the super-auto. Lets say Frank and Jen can make a coffee to the same quality and consistency. Why would I take the process away from Jen and hand it over to Frank?
There is only one way I would do this and it’s if Frank could make coffee 15% better than Jen. Now, if I’m not a speciality coffee shop, this wouldn’t be the case. As a restaurant or a conference centre, Frank offers me increased efficiency and less concern around staff training. But I do not represent those businesses. This article is really about the question of super-autos in speciality shops. *
The truth is Frank still requires Jen to program the recipe and check the result; ultimately the flavour reference will always be human. In a business where there is a constant house blend the human involvement would be very limited.
The claim about the best auto machines is that they can make coffee better than 95% of baristas. Although logically, to focus on improving your worst cup is sensible, no self-respecting speciality shop wants to be (or sees itself as) outside that top 5%. Would the shop actively choose a tech that doesn’t allow them to achieve the upper echelons of coffee perfection?
But, for arguments sake lets say there is an auto machine can achieve the top 5%. What then? I save staff time and efficiency and money… right? Well every change has pros and cons, so the more pertinent question here is, what do I lose?
It is easy to use numbers as our metric for success. They present a nice reference point that feels concrete. But are they actually back and white? Often a number is an average and an approximation.
At its core coffee is subjective and experiential, and a lot of value in coffee comes from everything around it, it is the processes that have gone into it and the perceived value. This is often an uncomfortable notion in coffee. We want great coffee to speak for itself, but often it can’t. It needs presentation and context. It’s important to understand how people perceive coffee quality and where their perception comes from, rather than to ignore it and believe that the coffee will do the job all on its own, just because that idea sits more pleasantly with our ideals.
Ironically a coffee shop is one of the hardest places to focus on provenance, flavour and character. I mean, sure, an office block is even harder, but a cafe is tricky as the primary use of the cafe is not a journey into the flavour and complexity of coffee. Most coffee shops are limited to their local audience and many of the customers are using the space to meet friends, have a work meeting or read a book. At best, most customers are after a ‘good’ coffee. There are exceptions with destination shops.
‘So why do you make capsules then?’ Situations when automation does add value.
When talking about automation and my scepticism of its role in speciality coffee shops, the comment, ‘coming from the guy who put speciality coffee in capsules’ has popped up more than once. The thing is, our exploration of capsules was never meant to take away the role of the barista, it is meant to do the barista’s role where they don’t currently exist. For me capsules are a great example of the real benefits of using automation in a really positive way.
Capsules don’t just offer convenience. It is often missed that in offering convenience they also enrich the drinking experience and open up choice for the drinker. It would be highly uncommon for a customer to buy three bags of beans and weigh them all out at the same time and brew three Aeropress side-by-side to taste the difference in character.
With capsules a customer can, of their own accord, explore the range at a push of a button. Routinely, customers surprise themselves with how varied the flavours are, and their ability to pick up on them. At home is actually potentially a better place to present speciality coffee, somewhere that the individual takes time and focuses on the special ingredient they bought without distraction.
A shop needs to create an experience to help focus visitors on the coffee. Take a look at the big chains, Many of them made the move from manual to automated years ago and now with their premium stores they are moving back to manual processes. At trade shows they discuss with various companies that they are looking for theatre, provenance and process to engage the visitors.
A customer’s experience is dynamic. Speciality coffee should have cup quality at its centre but we should understand how to furnish this. The truth is that we already do understand, which is why the move to automation will be so unlikely. We know that customers associate automation with fast, convenience non-premium coffee experiences. Making the change would be to fly in the face of this assumption and it would require a lot of work to tell customers that you haven’t sold out but are going auto for a good reason.
We took part in a latte art sensory study with a sensory scientist called Charles Spence. He is responsible for much of the focus on sensory science in gastronomy over the past decade. The question was simple: if we make the same cup of coffee twice, but put latte art on one and not on the other, will people perceive the latte art coffee as better?
The result was interesting. The participants did not think there was any difference in quality; they noticed that the coffees were the same. The interesting part was that despite this they were willing to pay significantly more for the time, energy and human element that went into the cup with latte art.
It’s often mentioned that coffee is going through its industrial revolution, we have seen it with so many other products, and now it is coffee’s turn. But it’s wrong to think of coffee like a pair of trousers, it’s better to think of it as a culinary product. Throughout food and drink, much automation takes place but it never takes over the boutique upper echelon, mostly for the reasons explored here.
The trousers are not made in front of you. The way in which they were made is hard to identify. In a coffee shop we see the coffee freshly prepared in front of us. This is why I think speciality coffee shops will never go super-auto. Much of coffee’s story and quality is presented visually. It is this immediacy of preparation that urges shops to stay relatively manual, especially when combined with the fact that the super-auto will only do the job just as well, maybe.
Hasta la vista barista, the argument for automation.
There is a widely held viewpoint around super-autos that argues the opposite. It has been pointed out that by taking the physical job away from the barista, we allow them to flourish in other ways. The barista can focus on customer experience and the shop can attain a more immersive environment and expansive coffee range.
I don’t really understand this argument. I don’t think that a shop’s service or offering is really defined by coffee-making technology at all.
Service is an inherent company culture based around people and ethos and the way you intend to engage with customers. If anything, I think that the disengagement of not being physically involved in making the coffee potentially leads to less engaged and potentially less passionate baristas. Sure, it’s hard to make coffee and talk to people all the time, but this can be managed in various ways, and losing the making will in turn lose a major focus at the centre of speciality coffee shops.
At the same time freeing barista’s time up to focus on more of a sommelier like experience is not a natural evolution. It is instead a dramatic change in the way the shop presents itself to customers and this would in turn require a big change from customers in terms of the engagement they are looking for.
The offering in most coffee shops of a house coffee and a guest coffee isn’t really defined by how the coffee is being made. It is largely defined by the experience the business has decided to offer its customers based on what it thinks they are after (and what they want to offer). There is a clear way coffee tech can broaden this potential ‘list’ of coffees, but the challenge they will face is whether their customers are actually interested in this list.
In reality a super-auto in most shops will just mean the following experience:
Barista: Hello, how are you?
Customer: Good thanks, can I have a flat white please?
Barista: Absolutely, please take a seat.
Barista presses button.
Coffee makes itself and next customer approaches the counter.
The opportunity to create, intrigue and to engage becomes harder. Now it is wholly reliant on verbal communication with no use of visual stimulus, which is often more pervasive and impactful. Of course, there are ways visuals could be designed to help in which the coffee is no longer doing the job.
And there will be exceptions. I can imagine a fascinating concept model with super-autos and a fine dining style service experience.
I expect to see more automation throughout coffee. But it I predict it will be dynamic and in the areas where it already adds value, which mostly isn’t speciality coffee shops. Even then the impact of automation is often overplayed. For example, the idea that automation will make airport lounge coffee great is misled. It comes down to looking at what is most impactful on cup quality. The main reason the coffee I had in the airport lounge on the way back to the UK this past week was pretty dreadful was not because of the brewing technology (as lacking as it was) it was because of the quality of coffee beans being used.
Predicting the future.
In summary, my prediction is that the march of the automated machine will stop on the doorstep of speciality coffee, that we will see more automation across the industry and we will see speciality producers embracing automation in a way they haven’t before, but the landscape in speciality shops will remain a mix of human and manual processes combined with new tech and ideas.
Then again most predictions are wrong. Last week at the WBC someone suggested that I set a reminder for five years from now, so I can look back on this and see how right or wrong I was. I think I might just do that. See you in five years Jen and Frank.
*. To clarify, to me the 15% better idea is a fairy tale discussion. There is a belief in coffee that we are about to get substantially better. The reality is that we make marginal gains at the top, often we are gaining more control over our desired output, but the best cup of coffee in the world will not get 15% better for no obvious reason, and certainly not because we just handed it over to a machine. The machine will just do the job we ask it to. In reality machines are playing the catch up game, and yes it’s plausible they will get to the same level. But again, what does this mean?