This blog first appeared as a series for the Raw Material newsletter and has been slightly altered for publishing here.

Quality Control. It’s intricately built into the fabric of speciality coffee, a part of the daily routine for roasters, baristas, importers, exporters, farm managers… Afterall, how can we ensure the quality that speciality is built on if we don’t control it?

Putting aside the slightly ludicrous idea that we can somehow control the quality of a natural product that is prone to the whims of the ecology it grows in, let alone all the hands it passes through to get to the cup (frankly it’s a miracle coffee tastes so good so frequently!). The largest issue with QC is the binary nature of its application, approval or denial. What this leads to is very little in the way of a practical or effective feedback loop from consuming countries to producing countries, a hole that has been filled by social media it seems. 

But let’s narrow our focus a little, tear ourselves away from the wider view of the value stream and concentrate on the import/export of green coffee. For it is here that QC is perhaps most crucial to the sale of coffee. The physical and sensorial properties that are measured include, moisture content, water activity, density, SCA/CQI sensory score, defect count, UV analysis, and screen size.

These are then compared against contract expectations and export regulations, to ensure they meet specifications prior to shipping, and again on landing in the consuming country. Contracts can specify a number of qualities and/or sensory profiles, and this is where the binary nature shows itself. Based on these samples, contracts are either rejected, or approved.

This isn’t usually an issue for the producer or smallholder, as they have already been paid and the risk that comes with refusal has been passed downstream. But it also doesn’t help them much. All of this work at laboratories around the world actually has very little impact on either the quality of the coffee, or its viability as a saleable product.

Which raises a question: how (and when) can you give feedback to producers in a manner that will have a tangible impact on not only the coffee they produce, but also their livelihoods?

The answer is rarely and with a light touch. Any feedback given can have wide ranging effects, carry unintended consequences, and create risk for the producers. To enter into this discussion without this in mind perpetuates the power dynamics that are at the heart of the inequalities that plague the value stream. Too often have those downstream asked for special processes, changes in variety, changes in farming practice etc, without either the requisite knowledge or the assumption of the risk involved. But it’s not impossible…

Full disclosure here, I work in green coffee for Raw Material. As well as importing into Europe and Australia we work in several countries where we are partners in the upstream operations, notably Timor Leste, and Burundi. In both countries, our Belfast based lab acts as both the pre-ship and landed QC centre, and provides valuable and tangible feedback that has been able to have real impact. The best way to show this is to look at a pair of case studies.

Timor Leste:

Mare Mesa, RM TL, operates a number of washing stations in remote areas of Ermera in Timor Leste. Infrastructure is lacking as a rule, and the climate in recent years has made efficient processing a challenge, specifically in regards to drying. 

How the parchment is dried is crucial to ensuring both quality and shelf life. In 2022, we found that the coffees landed exceptionally better than the preship samples. However, a month later the washed coffees had not only faded, they had dropped off a cliff of quality. Complex acidity and sweetness had faded to papery and woody notes seemingly overnight. A tell tale pointer to why was revealed in the QC.

As a matter of course, we conduct a UV analysis. The number and state of the glowing beans can tell us things about the history of the coffee’s processing. For instance, bright spots often indicate pulper damage. In this case, the coffees were devoid of the bright glowing beans often associated with fermentation issues and often indicate a shorter shelf life, and instead exhibited an intense mottled glow across the whole bean surface, much like static on a TV screen.

This mottled glow can come from two major sources;

Intense heat during the initial drying phase and heat build up during the milling process.

The first was easier to diagnose and correct, it turns out it was an unusually hot year and the beans were being laid out to dry at the peak of the day with no shade. Naturals and Honey processed beans can handle this as they have a built in heatsink to protect the bean from this exposure, but the washed coffees in their parchment do not. This explains why the washed coffees exhibited faster fade than the naturals. The solution was a change in process and shade netting for the washed drying beds, practical infrastructre changes that were low risk and easy to implement.

The second reason was harder to diagnose, whilst training on the UV process at our farm in Colombia we micromilled some coffee for extended periods to see if there was any change in the UV. The longer periods created more heat and the beans exhibited identical edge glow levels to the beans from Timor.  An exploration of the milling revealed that for that harvest a different mill was used, a commercial mill that runs continuously and as such gets very hot and is not designed for speciality lots. The solution? Build our own mill… in the future, but in the meantime work with another mill.

Following seasons crops have landed with far lower edge glow and much longer shelf lives, the impact here is a growing market for Timorese coffee allowing it to be purchased more at sustainable prices.

Burundi:

In Burundi we were able to have a similar impact, but this time measuring water activity.

The Izuba Company, aka Raw Material Burundi, manages the Izuba coffee washing station in Kayanza in northern Burundi. Our partnership with Muraho Trading Company has precipitated a knowledge exchange that has resulted in well-processed and well-dried coffee. However, even with much caution, some coffees in the past lacked consistency across lots, and had a tendency to fade.

As part of the pre-ship and landing QC, we measure both the moisture content and the water activity. Washed coffees have been shown to have a predictable relationship between the two; and the levels of both can be used to predict potential issues with shelf life, amongst others potential problems such as mould.

In the lab, we measured differences between the lots with measurements ranging from 0.49 – 0.63 on the pre-ship samples. This range closed slightly on landing, likely due the nature of the residual effect of water activity readings, and the homogenisation of water content as the coffee matures. These differences can lead to disparities in the shelf life of the coffees. The coffees with the higher readings are more likely to fade faster.

Discussions with the station team lead to the discovery that at that time, they were not measuring either the moisture content, or the water activity during the drying process. They relied as many do on experience to tell them when the coffee was ready. This approach shows a level of experience that is admirable, but the opportunity was there to have an impact on the commercial viability of all of the coffee from the station. And by extension, impact the welfare of the smallholders, through increased purchasing potential from future crops, if we could stabilise the drying.

Water activity meters are expensive, and not easy to come by; particularly in one of the most economically challenged countries in the world. Thankfully, although psychrometers work in the opposite way, they are cheap, portable, and easy to use. Where water activity meters work by measuring the water vapour pressure in the sample, psychrometers measure relative humidity; which in our case with a sealed sample, is analogous to water activity. We use the same meter in the lab, so calibration between us and the station was simple.

Present crops have landed with a variance of only 0.2 – 0.3 and have aged excellently, showing positive qualities more than 12 months post harvest. In both cases, QC has been used not to refuse a coffee, or to speak negatively of its quality, but to make the best use of the feedback loop which creates lasting impact for the smallholders we purchase from.