Forecast design at GE Healthcare Europe: design of the
process for the unit production forecast

Abstract: Predicting customer demand is the key factor to achieve excellence in
supplychain performance (Rick Hoole & Shrunti Mandana, 2005). When
businesses become more and more competitive, excellence in the supply chain
turns into a prerequisite for survival. Making predictions of the future is
seldom easy and is dependent on gained experience from historical events –
events of either qualitative or quantitative character. A forecasting
process needs to be in place to point the direction where to seek inputs
and how to evaluate them making the most accurate prediction.

This paper is about designing a forecasting process for the creation of a
production unit forecast for the European region. Protein Separation - Sales
Support Europe (PS-SSE) is a support function within one of the businesses
of GE Healthcare and is responsible for the creation of the production unit
forecast for the European region.

In the paper the empirical study showed an existing process highly
dependent on qualitative inputs from senior forecasters due to lack of good-
quality inputs from quantitative sources. The process wasn't defined or
recorded which led to that experience from past judgments were lost and was
a reason to inconsistent usage of source data.

The paper has identified the need for tracking new sources such as net
demand, gross demand and prospect development. Together with budget data
these sources has shown to provide vital information in the forecasting
process. The paper provides information how the sources finds their way
into the forecast and also gives measures for evaluating their quality.

Simplicity, accuracy, traceability and stability have been pointed out as
important guidelines throughout the design of the forecasting process.
Understanding has been weighted to complexity and graphical presentation is
emphasized for the presentations. Scattered source data are brought
together so that cross-calculations between different source data can be
carried out. The calculated values together with other vital input are
lined-up and presented for the forecaster at one place.
Reviews are a cornerstone in a forecasting process and without them the
forecaster is blindfolded in his attempt improving the existing process. In
the paper suggestions for different areas of reviews been suggested, not
only for the final forecast and not always issued by PS-SSE.