Note that this is focused on espresso preparation. The paper and diagrams pair great with James' Espresso prep series.

James actually did a video on this paper specifically as well.

TL;DR: using a lower dose (15g) and a coarser grind produces a more complete extraction as well as a more consistent cup, allowing you to use up to 25% less material while producing a similar result to mainstream prep methods.

SUMMARY

Espresso is a beverage brewed using hot, high-pressure water forced through a bed of roasted coffee. Despite being one of the most widely consumed coffee formats, it is also the most susceptible to variation. We report a novel model, complimented by experiment, that is able to isolate the contributions of several brewing variables, thereby disentangling some of the sources of variation in espresso extraction. Under the key assumption of homogeneous flow through the coffee bed, a monotonic decrease in extraction yield with increasingly coarse grind settings is predicted. However, experimental measurements show a peak in the extraction yield versus grind setting relationship, with lower extraction yields at both very coarse and fine settings. This result strongly suggests that inhomogeneous flow is operative at fine grind settings, resulting in poor reproducibility and wasted raw material. With instruction from our model, we outline a procedure to eliminate these shortcomings.

EXCERPT - Systematic Reduction of Coffee Mass by Downdosing and Grinding Coarse

As we demonstrated in Figure 3, our model informs us that a reduction in dry coffee mass results in an increased EY max (shown schematically in blue in Figure 6). Thus, a barista is able to achieve highly reproducible espresso with the same EY as the 20 g espresso by reducing the coffee mass to 15 g and counter-intuitively grinding much coarser (as shown in red, Figure 6B). This modification may result in very fast shots (<15 s), a reduction in espresso concentration, and a different flavor profile.

The Specialty Coffee Association espresso parameters mandate that the extraction should take 20–30 s; we speculate that this might be partially responsible for the pre- vailing empirical truth that most coffee is brewed using grind settings that cause partially clogged/inhomogeneous flow. Remembering that the initial tasty point may lie in the clogged flow regime, some of the bed is extracted much more than the refractive index measurement suggests. By lowering the dry coffee mass and grinding to maximize EY, the operator may notice that they are able to push their extractions much higher than before, while achieving highly reproducible espresso.