HeyMark
Content

Blog

Weighted average: formula and a worked rate example

3 min read

A weighted average combines values by assigning a weight to each one. Multiply each value by its weight, add the products and divide by the sum of the weights. The result depends on both the values and what the weights represent.

Marketing reports often run into this problem when combining rates from campaigns with very different impression counts. A small campaign can have a high rate while representing only a tiny share of total activity.

Weighted average formula

Weighted average = sum of (value × weight) ÷ sum of weights.

Equal weights produce the usual arithmetic mean. When a weight represents a number of observations, the corresponding value must describe those observations. OpenStax develops the calculation in its explanation of the weighted mean.

A two-group advertising example

Suppose a report groups ads from the same platform and week. Both groups count the same kind of click, and their impression sets do not overlap. These numbers are fictional:

Group Clicks of the chosen type Impressions
A 10 100
B 90 9,000

Group A has a CTR of 10 ÷ 100 × 100 = 10%. Group B has 90 ÷ 9,000 × 100 = 1%. The simple average of those percentages is (10% + 1%) ÷ 2 = 5.5%. That gives the average of the two group rates, with equal weight for each group.

To find the number of clicks per 100 impressions across both groups, use impressions as weights:

(10% × 100 + 1% × 9,000) ÷ 9,100 ≈ 1.10%.

Check it directly from the totals: (10 + 90) ÷ (100 + 9,000) × 100 = 100 ÷ 9,100 × 100 ≈ 1.10%. The result is close to group B’s 1% because that group accounts for almost all impressions.

Weighted average calculator

Combine two groups. The example uses CTR values of 10 and 1, weighted by 100 and 9,000 impressions.

Calculation inputs

Enter a rate as 10 for 10%, without the symbol. Values must share a unit. Weights must represent the same base in both groups.

Initial example result

Weighted average
1.0989%
Simple mean of both values
5.5%
Sum of weights
9,100impressions

The result keeps the values’ unit. The sum of weights keeps its own unit. The simple mean gives each group equal weight.

Results are rounded to up to four decimal places; very small values use scientific notation.

Weights can be zero, but their sum must be positive. For CTR, use comparable impressions without duplicating them. For other data, keep units and definitions consistent. Each field accepts up to six decimal places and a maximum magnitude of 1,000,000,000.

The denominator determines the weight

For rates sharing the same definition, weighting by their denominators reproduces the overall rate. A rate of sessions with an action uses the corresponding sessions as its base. Cost per click divides total cost by the included clicks. Each group needs the same unit, period and definition.

Do not weight an impression-based CTR by followers. Avoid adding rows that already include a subtotal or repeat the same ad: duplicating its base changes its influence. If you need to establish what the click column includes, start with the CTR definition.

Volume weights and judgment weights

You can deliberately assign weights to evaluation criteria such as editorial quality and on-time delivery. That average summarizes priorities you chose. When reconstructing an overall rate, weights come from the observed bases. Mixing those purposes can produce a number that is hard to interpret.

For example, giving every client equal weight answers a different question from giving every impression equal weight. Write the question before choosing the calculation: an average across clients, or a rate across the combined activity.

Check the result before sharing it

Keep unrounded values until the final calculation. Distinguish a zero base from an unknown one. With zero impressions, that row’s rate is undefined; with unknown impressions, you cannot treat its weight as zero. The sum of the weights used must be greater than zero.

Download the example CSV to repeat the click and impression totals. Present those totals alongside the aggregate rate and keep group rates available, so the overall result preserves differences that may deserve investigation.

Plan your next post in HeyMark.

Keep the idea, review the draft with your team, and see how it performed in the accounts you connected.