ForecastQT™ is a web-based application used as a Forecast Performance Management system. It does not create forecasts. We think of it as an "engine management system," sitting on top of your forecasting engine and S&OP processes, providing your mechanics – or forecasters - with the tools to diagnose and fix problems, while learning how to recalibrate to avoid them in future.
ForecastQT™ imports forecast and actual demand data on a weekly/monthly cycle from whatever systems are used to produce them. It then uses proprietary methods to analyse forecast error:
- Accounting for the ease or difficulty of forecasting (known as 'forecastability').
- Separating avoidable error from unavoidable.
- Then decomposing it into bias and variation – (having different causes and effects on the business).
- Separating statistically significant signals (requiring action) from noise (requiring abstention).
- Calculating the impact of bias and variation on stock and costs, including the impact on customer service and potential lost sales.
- Then, by comparing performance to a consistently applied ‘do nothing’ benchmark, ForecastQT™ determines whether, where and when your forecasting is adding (or destroying) value and by how much.
Forecast performance can then be tracked across any dimension (e.g. product, business, channel) and at any level (e.g. category, family, SKU). To help accelerate acceptance of more sophisticated approaches, ForecastQT
™ also includes a full range of conventional metrics.
Most of the benefits that ForecastQT™ can deliver come from operational use in routine demand management and S&OP processes - perhaps focussing on eradicating bias (chronic over- or under-forecasting) first, as a quick route to value. However, it can also be used to provide ad hoc or periodic audits of forecast performance.