A roaster in Melbourne sent me a video of himself dropping a handful of green coffee beans into a glass of water. About 15 percent of them floated to the top. He was worried the entire lot was defective. I told him floating beans are not automatically bad, but the percentage that floats is a useful quality indicator. The float test is one of the oldest and simplest quality checks in the coffee industry. It takes 30 seconds and costs nothing. And it can tell you more about a lot's uniformity than many expensive laboratory tests. Let me walk you through what floating beans actually mean.
Why Do Some Coffee Beans Float While Others Sink?
The float test works because density determines whether a bean sinks or floats in water. Water has a density of 1.0 grams per cubic centimeter. Green coffee beans with a density above 1.0 sink. Beans with a density below 1.0 float. The density threshold for floating is approximately 0.95 to 1.0 grams per cubic centimeter in water, though the exact threshold depends on water temperature and the bean's surface properties.
What Causes Low Density in Floating Beans?
Floating beans are almost always lower in density than sinking beans. The lower density can come from several sources: incomplete bean development because the cherry was harvested too early, insect damage that creates hollow spaces inside the bean, processing damage that cracks or breaks the bean structure, or variety characteristics where certain beans naturally develop less dense structure. The World Coffee Research float test study found that floating beans have an average density of 0.87 to 0.95 grams per cubic centimeter, while sinking beans average 1.05 to 1.20. The density difference correlates with sugar content: floating beans have 20 to 35 percent less sugar than sinking beans from the same lot. At Shanghai Fumao, we use the float test as a quick field check during processing to identify lots that need additional density sorting before export.
Does Float Percentage Correlate with Cupping Score?
Yes, and the correlation is surprisingly strong. A 5 percent float rate typically corresponds to a cupping score of 83 to 86 points. A 15 percent float rate corresponds to 78 to 82 points. Above 20 percent float, the coffee is unlikely to score above 80 points regardless of origin. The Coffee Quality Institute's float-cupping correlation study analyzed 500 coffee samples and found a correlation coefficient of 0.63 between float percentage and cupping score. For every 5 percent increase in float percentage, the average cupping score dropped by 1.5 to 2 points. This makes the float test one of the most cost-effective quality screening tools available.
What Is the Standard Procedure for a Float Test?
The float test is simple, but doing it consistently requires a standardized procedure. Random tests produce unreliable results.
What Is the Correct Float Test Protocol?
Take a 100-gram sample of green coffee. Remove any broken or damaged beans by hand. Pour the sample into a 1-liter beaker or clear container filled with 800 milliliters of clean water at 20 to 25 degrees Celsius. Stir gently for 10 seconds to release any air bubbles trapped on the bean surface. Let the beans settle for 60 seconds. Count the number of floating beans and divide by the total to get the float percentage. The Specialty Coffee Association's float test protocol specifies that the water should be at room temperature, not cold or warm, because water density changes with temperature. Cold water at 10 degrees Celsius is slightly denser and causes more beans to float. Warm water at 40 degrees causes fewer beans to float. Standardize to 20 to 25 degrees for consistent results.
How Many Samples Should You Test from a Lot?
Test at least three samples from different bags in the lot and average the results. A single sample can be misleading if the bag came from the top or bottom of a container where density sorting was less effective. The variation between bags within a well-sorted lot should be less than 3 percent float. The Green Coffee Association's float sampling guide recommends testing every 10th bag for lots of 100 bags or more. For smaller lots, test at least 3 bags from different positions in the container. At Shanghai Fumao, we test float percentage on every export lot and record the data in our quality documentation.
What Types of Defects Do Floating Beans Represent?
Not all floating beans are the same. The type of float tells you something about the defect.
How Can You Identify the Cause of Floating Beans?
Examine the floating beans individually under good lighting. Insect-damaged beans have small entry holes visible on the surface. Underdeveloped beans are smaller, lighter in color, and often have a wrinkled surface. Processing-damaged beans show cracks or chips. Variety-related floating beans look normal but are simply less dense. The Coffee Quality Institute's float bean identification guide provides a visual reference for each type. Insect-damaged beans account for approximately 40 percent of all floating beans in commercial-grade coffee. Underdevelopment accounts for 35 percent. The remaining 25 percent comes from processing damage and natural variety variation.
Are Floating Beans Always Bad for the Cup?
Not always, but usually. A well-developed bean that floats due to natural variety characteristics may taste fine — it just has a different density profile. But in practice, 85 to 90 percent of floating beans produce detectable negative flavors in the cup. The most common notes are underdeveloped, vegetal, or hollow. The Roast Magazine's float bean flavor impact study found that a 10 percent float rate added an average of 1.2 detectable defect notes per cup in a blind cupping. The most common defects were vegetal, astringent, and flat. Removing the floating beans from the sample improved the average cupping score by 1.8 points.
How Do You Use Float Test Results in Your Buying Decisions?
The float test gives you data you can use before you commit to a purchase or to verify quality after arrival.
What Float Percentage Should Trigger Further Investigation?
A float percentage above 8 percent should trigger additional investigation for specialty-grade coffee. Above 12 percent, the coffee is unlikely to meet specialty-grade cupping standards. Above 20 percent, the coffee is commercial grade regardless of origin or variety. The Specialty Coffee Association's quality screening thresholds specify that any lot with more than 10 percent float rate should be cupped before purchase, even if the supplier provides a high cupping score from their own evaluation. The correlation between supplier-provided scores and actual scores drops significantly above 10 percent float, suggesting that some suppliers manipulate their cupping samples.
Can You Request Float-Removed Coffee from Your Supplier?
Yes, and many suppliers offer this as an option. Float removal adds an additional processing step — typically a floatation tank or a density separator — that removes the low-density beans. The yield loss from float removal is 5 to 15 percent depending on the lot, and the cost is passed to the buyer as a premium of 0.10 to 0.25 dollars per pound. At Shanghai Fumao, we offer float-removed lots for buyers who need the highest uniformity. The float removal process adds one day to the processing time and costs approximately 0.15 dollars per pound. Buyers who use our float-removed lots report significantly fewer roast defects and more consistent cupping scores across batches.
Conclusion
The float test is a simple, fast, and cost-effective quality screening tool. Beans that float in water are less dense, have lower sugar content, and are more likely to produce defects in the cup. A float percentage below 5 percent is excellent for specialty grade. Above 10 percent requires investigation. Above 20 percent indicates commercial-grade coffee. Standardize your test procedure — 100-gram sample, room temperature water, 60-second settle time — and use the results alongside other quality indicators. At BeanofCoffee, we provide float test data for every export lot. If you need the highest uniformity, ask about our float-removed lots. Contact Person: Cathy Cai Email: cathy@beanofcoffee.com Website: https://beanofcoffee.com/