1. PlanIQ
  2. Create a forecast action
  3. Understand quantiles

Quantiles are an advanced tool that you can use in your forecasts. They're used for areas of data that show uncertainty. 

PlanIQ provides default lower and upper quantiles that you can apply to your predictions. You can customize the defaults to suit your forecast.

Your feedback is important to us. To ensure you have a good content experience, we'd appreciate if you could tell us how you find the guidance on this page and if it addresses your questions. 

To take part, complete the quick survey at the end of the article, and tell us how we can improve. 

PlanIQ produces probabilistic forecasts. This means that PlanIQ algorithms produce a distribution of possible values, rather than a single point forecast. This distribution can be divided into quantiles.

In forecasts, quantiles help to address the uncertainty in forecasted values. Quantiles define a prediction interval within which the actual value is likely to fall with a given probability (P). 

For example: 

  • A P10 quantile indicates that the true observed value is expected to be lower than the forecasted value 10% of the time. P10 is a probability of 10%. 
  • A P90 quantile indicates that the true value is expected to be lower than the forecasted value 90% of the time. The difference between the P10 and P90 defines an interval range of 80%. 
  • This means that the probability that the true value falls between the forecasted values for the P10 and P90 quantiles is 80%. If you increase the difference between the quantiles, this would increase the interval range and the probability.

In the case of the median, or P50, 50% of the distribution falls on either side of the cut point. For quartiles, 25% of the distribution is in each interval at P25, P50, P75.

The selection of quantiles should align with your business considerations of the relative costs for over and under forecasting.

You can use the lower quantile, in instances where the cost of over-forecasting outweighs the cost of under-forecasting. 

For example: 

  • There are high costs related to overproduction, or too much stock. 
  • A manufacture has a high cost of capital, and a contact center whose objection is to save on labor costs.

The upper quantile can be used in cases where the cost of under-forecasting outweighs the cost of over-forecasting. 

For example: 

  • A retail unit has sufficient inventory space. The loss in sales due to being understocked outweighs the cost of being overstocked. 
  • It's useful to forecast at a higher quantile. 

A P50 quantile indicates the true value is expected to be lower than the forecasted value 50% of the time. This provides a balance between the concerns of over-forecasting and under-forecasting.

By default, PlanIQ uses P10, P50, and P90 quantiles.