To form an objective estimate, your team needs to get outside its own frame of reference.
The process requires five steps:
- Select a “reference class.” Find projects like yours that you can use for comparison. Example: If you help manage a chemical company that’s building an olefin plant using new technology, you may think your reference class should be existing olefin plants. But you’ll probably get better results by looking at plants that use the new technology, not necessarily plants producing olefin.
- Plot the distribution of outcomes. Arrange your reference class members as a distribution, showing the extremes, the median and any clusters. You may have to guesstimate by calculating the average outcome, as well as a measure of variability.
- Make an intuitive prediction of your project’s place in the distribution. Your gut feeling will probably be biased, so use the next two steps to arrive at a more accurate forecast.
- Assess the reliability of your prediction. You can foresee certain events more easily than others. Forecasting tomorrow’s weather is more reliable than predicting next year’s Super Bowl score. The goal is to estimate the correlation between the forecast and the actual outcome, expressed as a coefficient between 0 and 1, where 0 indicates no correlation and 1 indicates perfect correlation. Don’t try this at home; use a statistician.
- Correct your intuitive estimate. Yes, we have our biases, so your guesstimate made in Step 3 is probably way too optimistic. In this final step, adjust your estimate toward the average outcome of the reference class, based on your predictability analysis in Step 4. The less reliable the prediction, the more the estimate needs to be regressed toward the mean. Don’t be surprised if the adjustment is hefty.