How to Verify Net Content on Packaged Goods: The NetFill Sampling Method Explained

If you manufacture or package food, chemicals, or consumer goods, you know that regulators and retailers demand accurate net content. But how do you prove your filling process meets the legal average and minimum requirements without checking every single package? The answer lies in a structured sampling plan known as the NetFill method, based on international metrology standards. This article explains what it is, how it works, and how you can apply it immediately with a free tool.

What It Is

NetFill is a statistical sampling and decision procedure used to verify that the declared net content (the weight or volume printed on the label) of packaged products complies with legal metrology rules. It follows the principles of the OIML (International Organization of Legal Metrology) recommendations and ISO 90-2 (now largely harmonized with OIML R 87), which are adopted by national legal metrology authorities, including GB (Chinese national legal metrology) regulations. The method combines two independent checks: an average-dependent test (to ensure the average content of a batch is not below the declared quantity) and a minimum-dependent test (to ensure no single package falls below an unacceptable negative error limit).

How It Works: Formula and Steps

The NetFill method uses a sample of packages (typically n = 10, 20, or 50, depending on batch size). You measure the actual net content of each package and then apply two criteria.

Step 1: Define the legal limits.

• The Tolerable Negative Error (TNE) is the maximum allowed shortfall per package, expressed as a percentage of the declared quantity (e.g., 9% for 100–200 g, 4.5% for 300–500 g, per OIML tables).
• The Average Limit is that the sample mean must be at least equal to the declared quantity.

Step 2: Compute the sample mean and standard deviation.
Let x_i be the measured net content of package i, and n the sample size.
Mean = (sum of x_i) / n
Standard deviation (s) = sqrt[ sum((x_i - mean)^2) / (n - 1) ]

Step 3: Apply the average-dependent criterion.
The batch passes the average test if:
Mean >= Declared Quantity + (t s / sqrt(n))
where t is the Student's t-value for a 95% one-sided confidence level with (n-1) degrees of freedom (e.g., t = 1.833 for n = 10). This ensures that even with natural variation, the true average is not below the label.

Step 4: Apply the minimum-dependent criterion.
No individual package may have a net content less than (Declared Quantity - TNE). Also, the number of packages below (Declared Quantity - TNE) must be zero in the sample for a strict pass. Some protocols allow a small fraction, but OIML R 87 requires zero packages below the TNE in the sample.

Step 5: Decision.
If both criteria pass, the batch is accepted. If either fails, the batch is rejected or subject to 100% inspection.

A Worked Illustrative Example

Example data (illustrative only):
A factory declares 500 g per bag. The TNE for 500 g is 15 g (3% per OIML). You take a sample of n = 10 bags and measure:
505, 498, 502, 496, 510, 501, 499, 503, 497, 504 (grams).

Step 1: Compute the mean.
Sum = 5015, Mean = 501.5 g.

Step 2: Compute the standard deviation.
Deviations: 3.5, -3.5, 0.5, -5.5, 8.5, -0.5, -2.5, 1.5, -4.5, 2.5.
Squared: 12.25, 12.25, 0.25, 30.25, 72.25, 0.25, 6.25, 2.25, 20.25, 6.25.
Sum = 162.5. Variance = 162.5 / 9 = 18.06. s = 4.25 g.

Step 3: Average test.
t for n=10 (df=9) at 95% one-sided = 1.833.
Required mean = 500 + (1.833 4.25 / sqrt(10)) = 500 + (1.833 * 4.25 / 3.162) = 500 + 2.46 = 502.46 g.
Your mean is 501.5 g, which is less than 502.46 g. The average test fails.

Step 4: Minimum test.
Lower limit = 500 - 15 = 485 g. All measured values are above 485 g, so the minimum test passes.

Step 5: Decision.
Because the average test fails, the batch is rejected. The process needs adjustment (e.g., increase the fill target) to ensure the true average meets the declared quantity with confidence.

Common Pitfalls

• Using the sample mean alone without the t-correction: this leads to false acceptance when variation is high.
• Ignoring the TNE tables: different declared quantities have different legal negative errors; using a fixed percentage for all sizes is wrong.
• Sampling too few packages for large batches: OIML requires larger samples (e.g., n=50 for batches over 10,000 units).
• Not calibrating the scale: measurement error contaminates your verification.
• Assuming a normal distribution without checking: the t-test is robust, but extreme outliers can distort the standard deviation.

Get Started with a Free Tool

Manual calculation is error-prone, especially when you have multiple products and batch sizes. Use the free NetFill calculator at https://www.6sq.com/tools/netfill/ to enter your sample measurements and declared quantity. The tool automatically applies the OIML/ISO 90-2 rules, computes the t-value, and gives you an immediate pass/fail decision for both the average and minimum tests. Stop guessing—verify your net content with confidence and keep your production compliant.