A roast log is valuable when someone else can use it to reconstruct the batch and trace the cause when a result deviates from expectation. That requires pre-roast data, in-process data, and output data to be recorded and linked along the same timeline, rather than just saving a few scattered numbers.

A log needs to be able to reconstruct the batch

When a batch has a problem, the investigator needs to answer: how did this batch unfold from start to finish? A good roast log lets you answer that without relying on the operator's memory. If the log only records final temperature and total time, most of the progression information has already been lost before the investigation even begins.

Raw-material data and initial conditions

Before looking at the machine, you need to know the batch's raw material: bean type, lot, moisture if measured, and charge weight. Many deviations trace back to differences in raw material between lots rather than to machine operation. Without this data, an investigation can jump straight to suspecting the equipment when the real cause lies in the input.

Time-based progression data

The most important part of a roast log is the temperature-over-time sequence, including both ambient and bean temperature if measured, along with operating markers like first crack and points where power changed. A single final temperature doesn't reveal the path taken to get there, while most of the information about the cause of a deviation usually lies in that very path.

Output for checking consequences

After a batch finishes, you should record post-roast weight, loss ratio, bean color, and any sensory evaluation if available. Output data helps determine the severity of a deviation and provides a basis for comparison against previously acceptable batches.

Reading a deviation as a chain of causes

When investigating, it's worth asking in sequence: what changed first — raw material, machine settings, or environmental conditions; what changed next in the thermal progression; and does the output match that chain of changes through a known mechanism? Reading this way helps avoid picking an abnormal number as the cause simply because it's the most visible one on the chart.

Measurement uncertainty

Every sensor carries its own error and lag. Before concluding that a deviation in the log signals a real problem, it's worth considering that it could just be measurement noise or an artifact of sensor placement. If possible, compare against data from a second sensor or measurement method before drawing a final conclusion.

A good roast log needs to support tracing back

The real value of a roast log lies in helping the reader trace from an abnormal sign back to the root cause, rather than just recording that something happened. This requires data detailed enough, continuous enough over time, and stored in a way that can be looked up again when needed.

Practical conclusion

A roast log with high investigative value needs to connect raw-material data, time-based thermal progression data, and output data into one continuous chain. When a deviation occurs, it should be read in causal sequence — what changed first, what changed after — and measurement uncertainty should always be weighed before concluding on the final cause.

References

• Hernández, J.A. et al. (2007). Analysis of the heat and mass transfer during coffee batch roasting. Journal of Food Engineering, 78, 1141–1148.
• Batch-Scale Simulation of Heat and Mass Transfer of Coffee Roasting in Spouted Bed Roasters (2025). Beverages, 11, 162.
• Wilson, K. (2014). Coffee roasting acoustics. Journal of the Acoustical Society of America, 135, EL265–269.
• Córdoba, N. et al. (2020). Coffee extraction: A review of parameters and their influence on the physicochemical characteristics and flavour of coffee brews. Trends in Food Science & Technology, 96, 45–60.