Copy this worksheet for each batch
Give every batch an identifier, such as LotA-01. Keep the paper or digital worksheet beside the original bag and its supplier information. Fill the preparation fields before roasting, concise event notes during operation when practical, and the remaining fields after prescribed cooling. Never let documentation compete with watching the roaster.
| Field | What to record | Why it helps |
|---|---|---|
| Identity | Batch ID; supplier/lot; process; available harvest/arrival information | A comparison requires knowing whether the green material changed. |
| Setup | Exact roaster/region; manual version; stock or approved modified chamber; ambient conditions | Controls and sensor meaning depend on the setup. |
| Input | Green mass and scale resolution | Batch size belongs with the program, rather than inferred from a scoop. |
| Program | Automatic profile or documented setting sequence; each later adjustment with elapsed time | A final setting alone loses how you reached it. |
| Observed events | Color change; first isolated pop; sustained sequence; confidence; cooling start/reason | Event labels can otherwise hide uncertainty. |
| Output | Cooled roasted mass; spill or sample removals; calculated mass loss | Losses during handling can distort roast comparisons. |
| Storage | Container, location and opening dates | Later taste changes may include handling differences. |
| Cup | Tasting date/rest interval; method; dose; water/yield; grind; temperature; concise notes | A change in brewing can masquerade as a roast change. |
| Next experiment | One permitted change, predicted observation and decision rule | It makes the next batch answer a question. |
Calculate mass loss and keep its limits visible
Mass loss % = (green mass − cooled roasted mass) ÷ green mass × 100. With an invented 120.0 g input and 102.0 g output, the loss is 18.0 ÷ 120.0 × 100 = 15.0%. Roasted yield is 102.0 ÷ 120.0 = 85.0%, which is useful for capacity and cost planning.
Sweet Maria’s explanation uses this measurement to support repeatability, while warning that different initial moisture contents undermine universal roast-level comparisons. Keep every bean when weighing, use the same cooling and weighing procedure, and record sampling or spills. The same percentage from two different coffees does not prove the same development or flavor.
| Fictional measurement | Calculation | What to conclude |
|---|---|---|
| 120.0 g green → 102.0 g collected | 15.0% apparent mass loss | A useful record if collection was complete. |
| Same roast, 2.0 g accidentally spilled before weighing | 120.0 g → 100.0 g; apparent loss 16.7% | Handling inflated the figure; do not call it a darker roast. |
| Another lot also loses 15.0% | Equal percentage | Compare its own observations and cup; equal loss is not sensory equivalence. |
Evidence: Sweet Maria’s: How to Calculate Weight Loss in Coffee Roasting, 17 August 2023
Describe development time without making it a quality score
An optional log calculation is elapsed time from your recorded sustained first-crack onset to cooling start. If those fictional events occur at 7:40 and 9:10, the interval is 1:30. Dividing 90 seconds by the 550-second total gives about 16.4%. This describes the labels you recorded; it is not a recommended roast target.
If first crack was masked or ambiguous, leave the interval uncalculated or mark a range. An exact-looking percentage based on a guessed onset provides less information than an honest note. Also preserve the full program: two batches can share an interval and percentage while having different earlier heating histories.
Taste at recorded intervals, with consistent storage
After prescribed cooling, choose packaging intended for roasted coffee and label it. Sweet Maria’s storage guide discusses containers and valve bags, light, heat, moisture and oxidation. A container should not turn a careful roast comparison into an uncontrolled storage experiment.
Choose two practical tasting dates and record the elapsed rest for each batch. For example, compare the same batch on days 2 and 5; these are observation dates, not universal peak recommendations. Use the same brew method and ratio, note necessary grind changes, and record whether the coffee became easier to brew or more enjoyable. Compare batches at matched rest intervals when possible.
If roasting on different dates makes matched ages impossible, write the mismatch rather than pretending it disappeared. Keep the storage method consistent and consult coffee storage guidance when portioning or freezing. Rest, storage and brewing method interact; there is no single calendar day that settles every roast.
Evidence: Sweet Maria’s: Product Guide—Roasted & Green Coffee Storage, 1 June 2024
Make the next two batches answer one question
Begin with a repeat of a completed batch before trying to improve it. If repeated baseline batches differ strongly, resolve changes in coffee, mass, installation, settings and measurement first. A comparison is easier to interpret when normal variation is smaller than the effect you want to examine.
- State a question in taste terms: “Do I prefer the sweetness and aftertaste from this permitted finish adjustment?”
- Keep the same green lot, allowed batch mass, equipment and preparation. Choose one adjustment the manual explicitly permits.
- Write the proposed change before roasting and record what you actually did; do not rename an unexpected result afterward.
- Compare with a consistent written brew recipe and recorded rest interval. A helper can hide the batch labels if you want to reduce expectation bias.
- Record concrete sensations and preference using the tasting guide or brew journal.
- Repeat a promising comparison before treating it as established. If results disagree, keep both observations and investigate the difference.
Sources & limitations
This guide combines source-based explanations and practical starting points. Coffee Jacob has not performed hands-on product testing or controlled tasting trials for this guide. Source check: 13 October 2026.