Roasting Mistakes

Coffee and Tea Roasting Mistakes: Yield, Moisture, and Batch Errors to Catch

The mistakes that quietly bleed margin in coffee, tea, and dry goods processing, each with its symptom, root cause, and the number that exposes it.

The most expensive error in roasting is treating roast loss as a fixed percentage. Symptom: your Roast Batch Yield reconciliation drifts 2 to 4 points batch to batch and green inventory never ties out. Root cause: moisture loss is not constant. A light City roast at 400F drops roughly 12 to 14 percent weight, a Full City around 15 to 16 percent, and a dark Vienna 17 to 20 percent, driven mostly by free and bound water plus organic pyrolysis. Fix: log drop temperature and time per batch, then set yield expectations per roast level instead of one blanket 15 percent. That alone tightens green usage forecasts by 3 to 5 percent.

Confusing wet basis and dry basis moisture wrecks Moisture Loss calculations. Symptom: your calculated water removed disagrees with the scale by 1 to 2 percent of batch weight. Root cause: green coffee at 10.5 to 11.5 percent moisture is quoted wet basis, but many drying references use dry basis, and the two diverge fast above 10 percent. A 100 kg green batch losing water from 11 percent to 2 percent removes about 9.2 kg wet basis, not the 9.0 kg a naive subtraction suggests. Fix: pick one basis, state it on the sheet, and confirm with an oven or NIR reading rather than trusting the supplier certificate.

Rushing green coffee straight into the roaster without a moisture check causes uneven roasts and rejected batches. Symptom: scattered color readings, Agtron variance above 5 points, and Scrap and Rework Cost spikes. Root cause: incoming lots vary from 9 to 13 percent moisture, and wetter beans need 30 to 60 more seconds and higher charge temperature to hit the same development. Fix: measure moisture on every lot at intake and adjust charge temp roughly 2 to 3F per point of moisture above your baseline. Catching one full-batch reject on a 50 lb roaster saves 30 to 45 dollars of green plus an hour of labor.

Sizing Grinder Throughput off nameplate rather than sustained rate starves the packaging line. Symptom: the Packaging Line Output number looks fine on paper but the line idles 15 to 20 percent of the shift waiting for ground product. Root cause: a burr grinder rated at 500 lb per hour delivers that only on coarse settings and cool burrs. Drop to espresso fineness and real throughput falls to 250 to 300 lb per hour, and heat buildup forces cooldown stops. Fix: time-study three settings across a full 8 hour shift and plan the line to the slowest grind you actually run, not the spec sheet.

Underestimating changeover time between products is the quiet killer of small-batch operations. Symptom: the schedule assumes 15 minute changeovers but the floor logs 40 to 90 minutes, and OEE availability sits 10 points below plan. Root cause: allergen and flavor carryover from a naturally flavored or oily dark roast demands a full teardown, not a blow-down. Fix: use Changeover Cleaning Time to split changeovers into wet-clean versus dry-swap categories. A dry roast-to-roast swap may run 12 minutes, but a hazelnut-flavored batch to a clean single origin needs 45 to 75 minutes of scrub and verify to avoid cross-contamination claims.

Flavoring dosed by feel instead of by ratio ruins consistency and inflates cost. Symptom: batches taste hot or flat and your Flavoring Usage per pound swings 20 percent. Root cause: standard coffee flavoring runs 2 to 3 percent by weight, so a 40 lb batch needs 0.8 to 1.2 lb, but operators eyeball the pour and drift. Fix: dose by scale to a fixed percent and log actual against target every batch. At 18 to 25 dollars per pound for flavor oil, a 0.3 percent overpour on a 40 lb batch wastes about 0.5 to 0.7 dollars each time, which compounds into thousands across a year of production.

Ignoring energy variance per batch hides a real cost and a process drift. Symptom: your gas or electric bill outruns the Energy Per Batch estimate by 15 to 30 percent. Root cause: cold-start roasts, low batch fill, and long between-batch idling burn fuel with no yield. A drum roaster idling between batches can waste 20 to 40 percent of the fuel a full roast uses. Fix: batch back to back to keep the drum hot, load to at least 70 to 80 percent of rated capacity, and track BTU or kWh per finished pound so a rising number flags a leaking seal or worn burner early.

Building Finished Goods Cost on green weight instead of packaged weight understates true cost per unit. Symptom: your quoted cost per 12 oz bag looks competitive but gross margin lands 4 to 6 points under forecast. Root cause: 100 lb of green yields only about 84 lb roasted, and grinding, packaging giveaway, and scrap trim another 2 to 4 percent, so cost per sellable pound is meaningfully higher than green cost divided by green pounds. Fix: run cost on net packaged pounds and fold Labor Per Pound plus scrap into the base. Overfilling bags by 3 percent on a 12 oz target alone gives away roughly 0.36 oz per bag.

Published 2026-07-01.