Two lots dried to the same 11% moisture. Months later, one gave off a mouldy smell and the other was fine. The moisture meter had shown the same number for both. Drying looks like "a process that just removes water", and it is not. Trust the single number on the moisture meter and you fall into this trap. We decide the end point of drying not by instinct but by two numbers: moisture, and water activity.
Moisture content is the proportion of water remaining in the bean. We bring it down to around 11%. Too high and mould and deterioration advance; too low and the bean shrinks and the taste thins. Just as important is evening the number out across the lot. If it varies, colour and aroma vary when roasted. We never skip turning the beds, for the sake of this evenness. Take a single reading with the meter and you miss the damp beans at the bottom of the pile. So we measure at least three times, in different places.
What separated the fate of those two lots was water activity (aw). aw is a measure of how much water is available to microbes; at the same moisture content, a higher value means mould and oxidation advance more easily. Accelerated storage tests have shown that the shelf life of green coffee changes with processing method (dry, honey) and storage temperature, and that the longer water activity stays high in the background, the higher the risk of mycotoxins (Saengrayap et al., 2024). Deterioration of green coffee is driven mainly by lipid oxidation: acid value, peroxide value and TBARS (thiobarbituric-acid-reactive substances) rise while moisture and unsaturated fatty acids fall (Chi et al., 2020). So drying is also the process that holds aw low and closes the door to deterioration. This is why we look at a water activity meter as well as a moisture meter.
Speed changes the taste too. Dry fast for relatively clean, fresh sweetness; dry slowly for the weight of a red wine. The long range of one and a half to two and a half months for naturals exists because that time itself is a point of flavour-making. But slow is next door to danger. The longer drying takes, the longer water activity stays high and the greater the risk of mould. Drying too fast, on the other hand, dries the surface first, opens a gap with the interior, and the beans crack or case-harden so the water inside never fully leaves. So drying is neither "all at once" nor "dragging on"; the key is moving water out at a steady pace.
We do not end drying on "it feels dry enough". When moisture approaches the target we measure, look at water activity too, and stop when both line up. The slower we want to go in a given year, the more we turn the beds every day and close the racks when rain comes.
"You cannot settle whether drying went well with your nose and fingertips alone. In the end it is closed by two numbers: moisture and water activity."
For example, this season's Geisha Honey was stopped after 24 days of drying at 10.8% moisture and 0.57 water activity.
The measured values of our lots are listed on the DATA BANK page. For the thinking behind keeping numbers, see What is the Data Bank; for the fermentation that sets the direction of taste before drying, Managing fermentation by numbers.
New records and lot releases, delivered to your inbox first.
New lots, drying updates and field notes — every Friday, in one short letter.