How Defect Counting Works in Coffee Grading: Sample Size, Equivalencies, and Buyer QC

2026 cluster role: This page owns the narrow defect-counting procedure. It does not own the broad Fine Robusta grading query. For the complete buyer framework, use Fine Robusta Grading: What Buyers Can Verify Before Cupping. For defect interpretation, use Physical Defects in Fine Robusta Grading.

Defect counting looks simple: take a green coffee sample, separate defects, and count them.

In professional grading, the result is only meaningful when the method is named.

Different systems can use different sample sizes, defect categories, equivalency rules, and acceptance thresholds. Therefore, “this coffee has three defects” is not enough information for a buyer.

A complete statement is closer to: “This lot was evaluated on this date, using this sample size and this named protocol, with these defect categories and equivalency rules.”

What defect counting is trying to measure

Defect counting provides a structured description of the physical preparation of a coffee lot.

It can help identify:

• serious damaged beans;
• immature material;
• insect damage;
• broken or chipped beans;
• foreign material;
• sorting inconsistency;
• preparation changes between lots.

The count does not directly measure flavor.

It is one layer of evidence.

Start with sample identity

Before counting anything, record:

• supplier;
• producer/group if known;
• country/region;
• harvest;
• process;
• lot code;
• sample date;
• sample source.

A defect report that cannot be connected to the offered or delivered lot has limited procurement value.

If the sample was taken from a larger shipment, note how it was obtained.

Representativeness matters as much as counting accuracy.

Sample size is part of the result

A defect count must be tied to sample weight.

This is important because articles often mix numbers from different systems.

A buyer should never assume a defect threshold from one sample basis can be copied directly onto another.

When OCC references a historical or current standard, state the sample basis used by that standard.

When OCC uses an internal QC sample, label it as internal.

The method should travel with the number.

Defect categories are defined by the standard

A grader does not invent categories while sorting.

The named standard defines how defects are classified.

Categories may include examples such as:

• black beans;
• sour beans;
• insect damage;
• foreign material;
• broken beans;
• immature beans;
• shells;
• other defects.

The exact classification and severity weighting depend on the system.

This is why a generic web article should not publish a universal equivalency table without the source.

What is defect equivalency?

Some grading systems do not count every damaged bean as one full defect.

Instead, several pieces or instances of a less severe defect may equal one “full defect” for grading purposes.

More severe defects may carry a different weighting.

This is called an equivalency system.

The objective is to account for differences in likely quality impact.

For buyer reporting, record both:

• raw observed count;
• calculated equivalent count where the named protocol requires it.

That makes the report easier to audit.

Primary versus secondary defects

Many historical and commercial systems group more severe defects separately from less severe defects.

The practical value is prioritization.

A small number of severe defects can represent a different risk profile from the same raw number of minor preparation issues.

However, the terms “primary” and “secondary” should always be tied to the grading system being used.

Do not assume every market defines them identically.

A basic defect-counting workflow

Step 1: prepare the workspace

Use consistent lighting, a clean surface, suitable trays, a scale, and the relevant defect reference materials.

Step 2: confirm sample weight

Weigh the sample according to the selected protocol.

Step 3: spread the coffee

Create enough visual separation to inspect the beans efficiently.

Step 4: remove foreign material

Record stones, sticks, metal, or other non-coffee matter as the protocol requires.

Step 5: sort defects by category

Keep categories separate instead of building one mixed “bad bean” pile.

Step 6: count the observed units

Record the raw count.

Step 7: apply equivalencies

Use only the equivalencies defined by the named standard.

Step 8: record supporting physical observations

Moisture, size distribution, odor, color, and other checks may be recorded separately.

Step 9: preserve the report

Attach it to the lot and evaluation date.

Why photos help

Photo records can support training and dispute resolution.

A photo can show:

• defect category examples;
• severity;
• unusual contamination;
• foreign material;
• overall sample preparation.

Photos do not replace the formal count, but they provide context.

For repeat suppliers, they can also show whether preparation is improving or deteriorating.

How to reduce grader inconsistency

Physical grading involves trained judgment, especially when a bean is borderline between categories.

Consistency improves when teams use:

• the same standard;
• defect reference samples/images;
• calibration sessions;
• consistent lighting;
• agreed category definitions;
• periodic cross-checks.

If two graders classify the same sample very differently, review the categories before averaging the results.

A disagreement can reveal a training problem.

Do not mix green defects with roasted quakers

Green defect counting and roasted quaker evaluation are related but different activities.

Some immature or problematic material becomes more visible after roasting.

Record roasted-sample findings separately.

This makes it possible to determine whether the observation was made in green coffee or after sample roasting.

Moisture is not part of the defect count

Moisture may be measured during physical grading, but it is a separate data point.

Do not convert moisture problems into an invented “defect count.”

Record:

• instrument;
• result;
• date;
• sample.

The same applies to water activity where used.

A clean defect count does not guarantee safe or stable storage if moisture conditions are poor.

Screen size is not a defect count either

Screen distribution describes size.

It can help a roaster understand physical uniformity.

It does not automatically increase or decrease the defect count unless the named commercial specification explicitly includes size tolerance.

Keep the categories separate.

A buyer report is more useful when each measurement has its own field.

Defect count versus sensory defect

A physical defect is observed in the sample.

A sensory defect is perceived in the cup.

They can be related, but they are not interchangeable.

For example, a small physical issue may have little detectable sensory effect in one sample, while another defect may create a clear taint.

Cupping verifies the sensory consequence.

For the distinction, use Grading vs Cupping in Coffee Evaluation.

Current Fine Robusta context

CQI’s current Q Coffee sample process says that Robusta samples are evaluated according to CQI Fine Robusta Standards and Protocols and are assigned to three certified Q Graders.

CQI’s current workflow should be distinguished from the professional Q Grader education/licensing program, which moved to SCA management on October 1, 2025.

Official references:

• CQI Q Coffee sample process
• SCA Q Grader program update

OCC should therefore date historical defect tables and avoid implying that one old web summary is the timeless current standard.

Why exact thresholds are easy to misuse

A threshold can change meaning when:

• sample size changes;
• defect categories change;
• equivalencies change;
• standard version changes;
• commercial contract uses a different specification.

The safest editorial rule is:

Never publish a Fine Robusta defect threshold without the source, method, and date context.

This protects both SEO authority and procurement accuracy.

Defect trends are more valuable than one count

One report is a snapshot.

Repeated reports create a quality trend.

Track by lot or harvest:

• severe defect equivalents;
• insect damage;
• broken beans;
• immature material;
• foreign matter;
• moisture;
• size distribution;
• sensory defects.

A supplier may improve even if one lot is not perfect.

Another supplier may show gradual deterioration that a single inspection would miss.

Trend data turns grading into supplier development.

Use defect data to investigate process stages

Physical results can guide questions.

High immature material

Ask about cherry selection, harvest timing, sorting, and flotation.

Increased broken beans

Ask about hulling, mechanical handling, storage, and transport.

Sour/black defects

Review fruit condition, fermentation, drying, and sorting.

Increased insect damage

Review farm pressure, sorting, and lot preparation.

Foreign material

Review cleaning, milling, bagging, and warehouse handling.

These are investigation directions, not absolute causal diagnoses.

How buyers should compare two suppliers

Do not compare only the final equivalent count.

Also compare:

• protocol;
• sample basis;
• lot size;
• harvest;
• processing;
• storage;
• repeatability;
• sensory result;
• delivery performance.

A slightly higher physical defect count may not be the most important commercial risk if another supplier is inconsistent in delivery or cup quality.

The quality system should support the purchasing decision, not replace it.

Acceptance ranges and contracts

A B2B buyer may define internal acceptance criteria in addition to formal grading systems.

That can include:

• physical preparation;
• moisture range;
• sensory profile;
• roast behavior;
• espresso extraction;
• packaging;
• lot variance;
• delivery condition.

If OCC uses an internal specification, label it clearly as an OCC or buyer acceptance range rather than presenting it as an official CQI or SCA requirement.

This distinction is essential for credibility.

Cambodia-specific use

Defect counting can help document improvements in Cambodian Fine Robusta preparation.

For example, repeated lot records can show whether selective picking, sorting, drying, and milling changes are reducing physical defects over time.

But do not convert one clean sample into the statement “Mondulkiri coffee has low defects.”

The evidence belongs to the measured lot.

For origin context, use Fine Robusta Cambodia.

A practical buyer worksheet

Identity

• lot code;
• producer/group;
• origin;
• harvest;
• process.

Method

• grading standard;
• version/date;
• sample size;
• evaluator;
• inspection date.

Raw observations

• defect category;
• raw count;
• foreign material;
• unusual physical notes.

Calculated result

• equivalency rule;
• equivalent count;
• pass/hold/review status.

Related measurements

• moisture;
• screen distribution;
• odor/color;
• quaker check.

Sensory follow-up

• cupping date;
• defect/taint notes;
• score/descriptive result;
• intended-use test.

This keeps the physical count connected to the larger QC record.

Common mistakes

Mistake 1: counting without recording sample weight

The result cannot be compared reliably.

Mistake 2: using raw bean count as equivalent defect count

Apply only the named standard’s equivalencies.

Mistake 3: mixing defects into one pile

Category information is lost.

Mistake 4: copying thresholds from an uncited article

Use the official or named standard.

Mistake 5: treating the count as a flavor score

Cup the coffee.

Mistake 6: assuming every grader classifies borderline beans identically

Calibration matters.

Mistake 7: losing the historical reports

Trend analysis disappears.

Mistake 8: applying the result to an entire origin

Keep it lot-specific.

Faq

How many beans equal one defect?

It depends on the defect type and grading standard. Use the equivalency table from the named protocol.

Is one damaged bean always one full defect?

No. Many systems use category-specific equivalencies.

What sample size should I use?

Use the sample size required by the protocol or commercial specification you are applying. Always record it.

Are defect counts enough to qualify Fine Robusta?

No. Physical preparation is only one part of the quality evaluation. Sensory evaluation and other requirements also matter.

Should foreign material be recorded separately?

Yes. It is a physical and operational risk and should be documented according to the grading method.

Can two graders get different counts?

Yes, particularly on borderline categories. Standard references and calibration reduce inconsistency.

Does zero visible defects mean the coffee will taste excellent?

No. Physical and sensory quality need separate verification.

Conclusion

Defect counting is valuable because it turns physical preparation into structured evidence.

But the number is only useful when the buyer knows the lot, sample size, grading standard, categories, equivalencies, and date.

For Fine Robusta, OCC should keep defect counting narrow and method-specific, then route the broad buyer question back to Fine Robusta Grading: What Buyers Can Verify Before Cupping.

For interpretation of specific physical defects, continue to Physical Defects in Fine Robusta Grading.