Simple Averaging of Forecasts

For example, if your team predicts quarterly sales will be $100K, $120K, and $140K, the average of $120K is likely closer to the actual outcome than any individual prediction. Try this: Ask at least three people for their independent forecasts, then calculate the mean of their numbers.

The mechanism relies on error cancellation—each forecaster brings unique knowledge, assumptions, and biases, so their errors tend to be uncorrelated and offset when combined.

Consider a product launch scenario: your marketing specialist predicts 5,000 units sold in the first month, your sales lead expects 8,000 units, and your finance analyst projects 6,500 units. The simple average of 6,500 units may be more accurate than any single guess because it incorporates diverse perspectives about customer behavior, market conditions, and historical data. Similarly, when planning a wedding budget, if three venue quotes are $8,000, $12,000, and $15,000, the average of $11,667 gives you a balanced expectation rather than relying on one potentially biased estimate.

The power of averaging increases with the diversity of your forecasters. In business settings, combining forecasts from sales, marketing, operations, and finance departments typically outperforms any single department's prediction. This happens because each team sees different pieces of the puzzle and makes different kinds of mistakes—some teams consistently underestimate, others overestimate, but these biases cancel in the average. Even better, this method requires no special expertise or complex calculations, making it accessible to anyone who can do basic arithmetic.

Simple averaging of forecasts is one of the most robust findings in prediction research, demonstrating that combined forecasts often improve accuracy over individual forecasts.

• Armstrong (2001): Meta-analyses show that combining forecasts from different methods often reduces errors compared to using single methods, and simple averages are frequently competitive with or better than more complex weighting schemes [1].

• Clemen (1989): Review of forecast combination research confirms that the "forecast combination principle"—that combined forecasts are generally more accurate than their components—holds across diverse domains from economics to weather prediction [2].

Simple averaging has important boundaries. It fails when forecasters share the same systematic biases—for example, if everyone anchors on the same misleading information, averaging will preserve rather than cancel that error. The method also struggles when there's a single expert with truly superior knowledge; averaging that expert with novices may dilute their insight. Additionally, social influence can also undermine effectiveness if forecasters see each other's predictions before making their own, reducing independence and causing errors to cluster rather than cancel.