Prufrocker: The Blog

Dialling-In Journal

Record scores, weights and measures and brew water information. Speed up the process of designing a recipe and identify which recipe is your overall favourite.

by  |   |  Coffee Science


Prufrock's dialling-in journal.

After tasting a range of shot times you will begin to develop a preference for one shot time over another. Take this a step further and record scores, weights and measures and brew water information. This will start to speed up the process of designing a recipe and will demonstrate which recipe is your overall favourite.

Above is a screen shot of the Prufrock Espresso journal on Google Drive.

Every morning the baristas at Prufrock have an hour to dial-in two espresso offerings, a decaf grinder and at least two brew bar options. If they prepare twenty different versions of each coffee to make sure they are targeting the best shot time every morning this would cost us a little over £3k a year in dry coffee alone and be a fairly scandalous waste of some special coffees.

If a coffee is new then we do need to run exhaustive tests to to design a recipe. We need to try a range of temperatures and shot times and brew formulas and this requires some investment and time. For the house blend that has been running for a couple of months they already know what recipe elicits the highest scores. Dialling-in is simply a question of getting onto the right grind setting. Successful recipes might need refinements as time passes and the green beans age, so when scores start dipping for a recipe, then further experimentation is needed. A good barista should experiment with any established protocols from time to time like a ‘devil’s advocate’, to make sure the existing standard really is optimum.

A scored espresso in the journal becomes a pivot-point for the rest of the day. Let’s say it poured in 33 seconds and was given a 3 for taste-balance. A barista can taste an espresso in the afternoon and see if it merits a higher score than the morning test shot. Maybe this shot poured in 28seconds and is resoundingly more balanced. Get the other baristas to taste it and reach a consensus on if it is an improvement. This can then become the target shot time until another shot comes along and outstrips it with an alternative recipe.

Note: At this stage in grinder development, opportunities to taste alternative shot times come along all too often, owing to inconsistencies in grind size distribution and the unwanted recalibration of grind settings as grinder burrs and casings heat up and expand. Friction from grinding will bring burr temperatures as high as 70deg Celsius during busy periods. We look forwards to discovering a grinder that is able to give us more stability.

Dialling-In App.

If in doubt, we look in the journal and see what was going on yesterday. The coffee is one day older. Do we need to grind finer? Will we need a longer shot time or a dash extra coffee to compensate for a little extra degassing of the grinds and a loss of aroma. Don’t know? Try it. Look back to last week in the journal. What did the barista do then and how did the flavour score compare to the day before?

We are trying to make our data more visible at Prufrock so we can take any parameter and pit it against another one. For instance, do we get better tactile scores from longer shot times? Put shot times on the vertical axis of a graph and your Tactile score on the horizontal axis and look for a trend. Maybe we want to work out just how long we should rest the coffee after roasting. We can put overall taste, tactile and flavour scores on one axis and the days from roast on the other. We plot the average score for the last three months all the 7 days from roast tests then, 8 days from roast etc. The highest peak will suggest an optimum resting period we can try to stagger our coffee ordering either side of to achieve the best overall tactile results.

One issue is that we find we often score quite similarly from one day to the next. At the moment the isolated events are what we are using to make conclusions from. It could be that the scale of 0-6 on the WBC score sheet needs to be more critical and we should look to give a percentage score. We could score Body, Round and Smooth separately rather than an overall tactile score. Similarly we are giving a taste balance score but not recording data on actual levels of perceived sweetness, acidity and bitterness from day to day. Please share any refinements you can suggest as we are only beginning this process.

Dialling-in philosophy

We are not wanting our baristas to express themselves with a style or recipe unique to them. Rather we want them to reach a consensus on what temp, dose and yield elicit the best overall scores. We trust our barista’s palates and we know they are skilled. They are creative and unique but this needs to be expressed through their interactions with the customers and each other, not with a trademark idiosyncratic style. With our recipe design we need to work like a jury to reach a unanimous decision on what parameters we are targeting.

Our equipment will not guarantee shot times. There is no reliable auto-pilot as yet so the baristas are the navigators. The farmer, roaster, the customers and the management are the cartographers. The farmer and the roaster have influenced what the coffee can taste like. Then the customer and the management have an expectation on what it should taste like. The management and the customers want a shot with runny thick crema that is really persistent. We want a super sweet, brightly acidic, subtly bitter taste-balance that reminds us of quality fruits and flowers and herbs. Or lovely caramel, nuts and chocolate and not any carbony flavours. We are looking for a full bodied, super smooth and really round espresso that persists in the palate for many minutes after it has been drunk.

It is difficult for a barista to deliver all this in the same cup. 24 seconds might give a great taste-balance with equal proportions of sweetness, acidity with a piquant bitterness. But it might not have a round and enduring tactile quality like a 33 second shot. Maybe 18 seconds delivers the most noticeable fruity, enzymatic flavours and these are muddied or muted in a 33 second shot. They need to reach a consensus on how to get the best overall score. And they need to taste all the shots that come through out of range of the existing recipe just to make sure it isn’t actually better. If the shot time target is 30 seconds, should you throw away a 24 second shot that comes through by accident just because the equipment lets you down. If it scores just as high as the 30second shots, you can establish a larger margin of error. If it is really a let down, you need to make a call on if you can justify throwing away shots on financial or ecological grounds.

The simplest way of dialling-in in a commercial environment is to target sweetness even for blends like Square Mile’s Sweet Shop that are not blended for the purpose of achieving balance. It tends to follow that a pleasant smooth texture should accompany an optimally extracted shot so sweetness is the starting point. Flavour is perhaps the easiest part, everyone always says they like the smell of coffee even if they don’t like the taste. There is plenty of truth in this as the aromatics seem to be released very early on in the extraction process. Targeting sweetness is the general direction to head in but at Prufrock, we are looking for the best overall scores. The better the map, the faster you will reach your destination.

All BRaT students that complete the Recipe Design Module at Prufrock will be given on going access to the Dialling-in journal.