A weighted average is appropriate whenever observations have different levels of importance or represent different-sized groups. Here is when it matters.
When weighted average ≠ simple average
Exam grades: a final exam worth 40% of your grade should not count the same as
a quiz worth 5%. Weighted average: (Quiz × 5% + Final × 40% + ...) / total weight.
Portfolio returns: if 80% of your portfolio is in fund A (returned 12%) and 20%
in fund B (returned 5%), your portfolio return is 0.8×12% + 0.2×5% = 10.6%,
not (12% + 5%) / 2 = 8.5%.
Customer satisfaction scores: if enterprise customers (30% of revenue) rate you
8/10 and SMB customers (70% of revenue) rate you 6/10, the revenue-weighted NPS
is 0.3×8 + 0.7×6 = 6.6, not (8+6)/2 = 7.
Average selling price across segments: a product sold at $200 to 100 customers
and $800 to 10 customers has an average ASP of (100×$200 + 10×$800) / 110 = $254,
not ($200 + $800) / 2 = $500.
When simple average is fine
Use a simple average when every observation carries equal importance and represents an equivalent "unit" — e.g. the average of 5 A/B test conversion rates where each test ran on the same sample size.
Use the weighted average calculator to compute weighted averages with automatic contribution percentages.