An Analysis of Nespresso

An Analysis of Nespresso

Update:5th August 2016 This was intended to be the first of a series of posts regarding Nespresso. Many people have asked for the follow-up, and I’m afraid that in the foreseeable future one isn’t coming. This post opened up a lot of learning, conversations, and discussion for me. However, some aspects of my advisory work make it difficult to continue to write on the subject as deeply as I intended. There may be more in the future but, in the short-term, there is unlikely to be a Part Two. 

This is the first in a little series of posts looking at how and why Nespresso works the way that it does. There’s a few things I think when it comes to Nespresso:

1). We continue to underestimate their success, and their ability to leverage technology to overcome hurdles of quality.

2). Speciality coffee roasters share more customers with Nespresso than they’d like to believe.

3). There’s a lot of speculation about the technology they use. Lots of it is clearly proprietary, so I wanted to dive into it a little bit more to try and understand what is happening.

So this is the first of the experiments (there’s more to come) and it is a pretty simple one. I wanted to look at how tolerant the pods were to variance in how user’s want to brew their espresso. This is a pretty simple experiment to conduct. I’m grateful for the assistance of Sang Ho Park in helping with some of the work.

For testing I purchased the cheapest machine directly from Nespresso: A Magimix Inissia in black. They’re painfully easy to use. Fill the removable tank with water. Plug it in. It heats up in under 30 seconds. You put a capsule in and push espresso and in 12 seconds you have an espresso.

The first experiment

I chose one particular capsule (in this case I chose Livanto) and brewed it at a variety of brew ratios, and measured the extraction.

Different Nespresso capsules actually have a range of dry coffee weight in them (I think from about 5.5g up to about 6.3g – but I haven’t tested them all). These particular capsules had 5.7g of ground coffee in them. This means you’re pretty much instantly going to brew at a totally different brew ratio to traditional espresso. I might favour a 1:2 ratio (e.g. 19g ground coffee, 38g liquid espresso). A Nespresso capsule works on a totally different flow rate, and I was at a 1:2 ratio in under 9 seconds!

I tested a range of beverage weights, from 15g of espresso liquid up to nearly 75g. Here’s a chart showing the extractions:

Screen Shot 2015-05-15 at 11.05.47

Now, I think (based on a few different factors) that what Nespresso is aiming for here is extractions above 20%. The pre-programmed espresso button produced a 30-31g shot that hit 21%. This is pretty impressive work for 12 seconds of brewing. If you’ve played with things like the EK-43 then your target extraction range probably moves from 18-22% of the Gold Cup standards, up towards maybe 20-24%. If this is your window then a Nespresso capsule hits that window regardless of where you pull it, between about 25g of liquid and about 60g of liquid.

The next thing to do was to test this with a capsule designed for their Lungo setting to see how this varies. For this second experiment I used two different capsules, both with a heavier dose of coffee – 6.2g.

DSCF7084

The preset button on the machine produces (quite repeatably) 95g of liquid in the cup. What surprised me was the difference in extraction between the two lungo capsules I used. The first clocked up at 25% extraction (The Vivalto capsule) while the Ethiopian lungo came in a little lower at around 21.49%. The second of these was much lighter roasted, and tasted distinctly like coffee from Ethiopia.

Absent-mindedly I decided to throw a lungo of the Ethiopian capsule (Bukeela ka Ethiopia) through a paper filter to see how it tasted without crema. This also seems like a good time to remind you what Nespresso have shared about crema. There’s a lot of speculation about how Nespresso achieves the crema it does. I have a few theories, but they’ll come later. For now, watch this video:

Back to my experiments: What surprised me was how quickly the coffee drained through the filter paper (if you’ve filtered a lot of espressos or lungos then you know this is weird). What surprised me even more was how few fines were present on the paper afterwards:

DSCF7089

So – you’re probably going to think the first thing I thought when I saw this: they’ve gotten rid of fines! There’s a lot of speculation about the grind profile in Nespresso capsules, and suddenly it looked like there was evidence of something interesting going on.

I can’t do much in the way of grind analysis, but we do have a few sieves at the roastery. These, I thought, would give us a little bit of insight into what was happening.

We took 50g of coffee ground for a lungo, and we shook it through the sieves we had. There’s limited data here but still interesting:

Trapped in the 500 micron+ sieve: 39.3g
Trapped in the 350 micron+ sieve: 4g
Trapped in the 150 micron+ sieve: 6.5g
Smaller than 150 microns: 0.2g

That’s clearly bimodal, and there are also clearly fines present. I then did what any sensible person would do: I put 12g of Nespresso coffee into a portafilter and pulled a lungo of a corresponding brew ratio. I then tested the extraction (it was higher – approximately 23%), I tasted it (worse than from the Nespresso machine) and then I poured the rest through a filter paper. The results were again interesting.

This is a filter paper after filtering a lungo using Nespresso grounds pulled through a portafilter:

This is a filter paper after filtering a lungo using Nespresso grounds brewed in a Nespresso machine:

So – a pretty notable difference. There are fines in the capsules, they just don’t end up in the cup when using a Nespresso machine. This leads us to two possible theories: the machine is filtering out the fines or the pod itself is filtering out the fines. This needs more exploration and this is going to happen in Part II of this blog post.

But wait…

What I haven’t really talked about here is how it tasted. I didn’t really like how the shots tasted, but I have a very different preference for many aspects of coffee and espresso compared to the typical (Nespresso) consumer. I’ve repeatedly tried to make the point that thinking we are somehow safe from the dominance of Nespresso, because we can make coffee taste better, is not a smart way to think. Should Nespresso want to make their coffee taste better – perhaps selling more than 6 billion capsules a year doesn’t feel like enough – then they face a technical challenge. No company is better positioned than Nespresso take on such a technical challenge, and while I’m not a fully paid up member of the Clayton Christensen school of “Disruptive Innovation” this does look like a pretty classic case.

They have a business model where they can buy whatever coffee they want, because they’re selling it at high prices. They’re selling their Ethiopian coffee at £53.23 per kilo (delivered of course…).

Am I trying to scare monger? No.

I am trying to pay a little more attention to a serious competitor. Here we have an option where you can put a capsule in a machine that is switched off, and in under 50 seconds have a shot of espresso better than most coffee shops around the world (accepting that most coffee shops are not good coffee shops). Our counter proposal: Spend hundreds, if not thousands of dollars, dial in a grinder for a while once the machine has spent 30 minutes getting up to temperature, then eventually pull a good shot. Drink it, then start cleaning up. Nespresso might be shockingly expensive, but so is a morning espresso pulled at home if it took a couple of goes to dial it in just right. Pulling shots of espresso is huge fun, if you want it to be. It’s a massive inconvenience if you don’t.

Speciality coffee doesn’t offer anything to the consumer who wants to drink great espresso at home, but doesn’t want a new hobby.

Whether or not I consider them a true competitor, or even if I don’t think they’re a threat to my business, there’s plenty to learn from them. I look forward to picking it all apart a little more and sharing it here in the future. If you’ve got questions or suggestions then let me know on twitter.

DSCF7067