Espresso Extraction

We roast our coffee to a relatively light degree which allows the origin characteristics of the bean to shine through. The down side of this approach is that we run the risk of underdeveloping the coffee, particularly when it comes to espresso. Underdeveloped coffee can taste harsh and sour, which are characteristics we definitely do not want to taste in the cup. Before releasing a coffee for sale we spend time roasting, tasting and measuring it, to make sure it meets our requirements.

One aspect we measure is the extraction yield of espresso shots. In simple terms extraction is the amount of coffee we have ‘squeezed’ out of the grinds and into the cup. For espresso we are trying to achieve an extraction yield of 18–20%. In other words, we’ll be throwing away around 80% of the grounds as the spent puck. If our roast is well developed we should be able to pull espresso shots which consistently achieve extraction yields in this range[1]. If our roast is underdeveloped we will fail to reach an extraction of 18%; the roasting has failed to sufficiently penetrate parts of the coffee bean.

We check for extraction yield by pulling a series of espresso shots which we then measure using a refractometer and a piece of software[2]. The refractometer provides us with a TDS (Total Dissolved Solids), or strength, reading and the software converts this into our extraction yield. The following is some of the data from today’s testing:

Coffee A (grey circle in image below)

Dose weight[3] Beverage weight[4] Extraction time[5] TDS Extraction yield
18.0g 37.0g 27 sec 7.4 15.76%
18.4g 37.0g 26 sec 8.0 16.67%
18.4g 36.5g 27 sec 8.6 17.68%
18.6g 35.0g 31 sec 9.0 17.55%

In each instance I failed to achieve an extraction in the target range of 18–20%. Even the longer extraction time for the last espresso failed to produce any favourable results. The first and third shots tasted harsh and sour, while the second was the best shot of the batch. As a result of these findings we returned to the roast profile for this coffee and have have looked at ways of developing it further.

Coffee B (blue circle in image below)

Dose weight Beverage weight Extraction time TDS Extraction yield
18.4g 38.0g 30 sec 8.8 18.83%
18.3g 36.7g 29 sec 9.3 19.33%
18.4g 36.1g 28 sec 9.5 18.99%
18.3g 35.3g 29 sec 9.6 19.19%

This time, in every instance, I achieved an extraction yield in the target range, despite being less than accurate with my beverage weights. The first shot, in particular, was an error as I didn’t press the button hard enough to stop the shot! However, this shot was also the tastiest on the table. Flavours of elderflower and orange we bursting from the cup, the espresso was sweet and left lingering tones of toffee. Possibly one of the nicest espresso shots I’ve tasted.

Espresso Extraction

Given the quality of the espresso shots, and the ease at which I achieved a good extraction yield, we’re happy with this roast profile.

Measuring extraction has opened up a whole new world for us as coffee roasters. When we’re trying to judge the quality of our roasting we need to be sure the espresso we are tasting has been made correctly. It has taken a couple of years to fully understand how to use a refractometer in our day to day coffee making, lots of research, plenty of reading and numerous failed attempts. It is a very powerful tool and one which we will continue to use, and learn from.


  1. the quality of roasting is only one of a number of variables which can affect extraction. Others include worn grinder burrs, overly fresh coffee, water quality, extraction temperature and barista skill.  ↩
  2. we use VST CoffeeTools on the iPad, which took a while to understand but is now paying us back for our perseverance.  ↩
  3. this is the weight of the dry grounds  ↩
  4. the weight of the resulting espresso; each espresso was pulled manually using scales  ↩
  5. typically aiming for an extraction time of 28–30 seconds based on the water quality at the Roastery  ↩

2 Comments Add yours

  1. Do you know how you compare with others in the industry. I can’t offer any evidence, but instinctively, I feel you are extracting longer shots and extracting for longer than a lot of coffee shops!

    I shall have to pay more attention to all the little numbers written on the grinders of all those speciality places I visit 🙂

    Thanks,
    Brian.

    1. Brian

      This was purely about checking roast development. We have very soft water so shorter shots can be a little inconsistent and I’m trying to remove all other variables. From here I’ll tweak the recipe, usually restricting the shot so my 18g dose will give around a 30g espresso. Likewise with the shot lengths – I tend to find a couple of seconds longer makes a difference up here with our water.

      I have a feeling a lot of places pull shorter shots because it is the fashion. An espresso normale would be twice the starting dose i.e. 18g to a 36g dose, so by that definition we’re about as ‘normale’ as us Celts can be!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s