 |
|||
|
GfK Healthcare | August 2010 | |
|
|||
Good Practices in Forecasting and Good Forecasting in Practice – Part II  By Rudiger Papsch, Managing Director, GfK HealthCare Asia The ability to make accurate forecasts is important for any organization, and this applies in particular to health care companies. Reliable forecasts allow market opportunities to be identified and optimally used, whereas poor forecasts can result in substantial financial losses through lost opportunities or misallocation of scarce resources. However, forecasting is a complex and challenging task and it is therefore not surprising that forecasts can be and often are erroneous. At the same time much progress has been made over the past decades in research on forecasting, directly tackling these challenges. Unfortunately these improvements are often not delivered to forecasters outside the academic community and are therefore rarely applied in practice. The two articles in this series are intended to bridge this gap by summarizing in 10 simple guidelines the advancements in forecasting. Last month’s Pipeline article addressed the technical challenges of forecasting through five principles that improve the accuracy of your forecasts. We describe these principles as ‘‘good practices in forecasting.’’ In this second part, we look at ‘‘good forecasting in practice’’ by sharing five principles that will help you deal with the organizational challenges of forecasting. Organizational challenges refer to the implementation of the forecasting process in the organization and the way the different stakeholders are involved in this process. The five principles presented here will both increase the accuracy of your forecasts and also ease your life by ensuring a smooth implementation of the whole forecasting process in your organization. Good forecasting in practice – how to tackle the organizational challenges 1. Keep it simple. The complexities involved in forecasting are not only a reason to prefer simple methods over complex ones, but also to keep the whole forecast and the underlying process of building the forecast clear and understandable. Besides increasing the accuracy of your forecasts this will also support the buyin of your stakeholders and facilitate their valuable input to your forecasts. 2. Make your forecasts transparent and involve the stakeholders openly and often. Involve all key stakeholders continuously in the forecasting process from the early stages onward. This will ensure their buyin and give you access to their domain knowledge. Be transparent by openly sharing the assumptions you make and methodologies you use. Be flexible and open to accept that the stakeholders might be able to contribute "better" market information than you have. Whenever involving stakeholders in the development of the forecast, however, it is key to discuss only the inputs to the forecast (assumptions) and not the outputs (results) because if stakeholders are allowed to directly influence the results of the forecast, they may tweak the results until the forecast predicts the outcome they prefer most. 3. Have a proper forecasting process in place including documentation and presentation. As mentioned, forecasting is a technical and complex process with many stakeholders involved. Forecasting therefore works much better if it is done in a systematic and organized fashion. Consequently, having a standardized forecasting process in place – and following it – is highly beneficial to your forecasts. Good documentation and presentation must be part of this process. Furthermore, a systematic approach and good documentation will make your forecasting process more efficient, as you will not have to reinvent the wheel with every forecast you do. 4. Manage the expectations of your internal clients and the management. The prediction of the future with precise numbers always involves a certain degree of error and risk of erroneous conclusions. Nevertheless, business decisions based on forecasts that follow good forecasting practices and methods will allow for much better and more successful decision making. Clearly communicate this fact to stakeholders. This rule applies in particular to the long-range forecasting usually needed in the pre-launch phase. The longer the time horizon the forecast has to cover, the less precise the forecast will be if only a single-point forecast is accepted as result (as opposed to forecasting a range in which the target variable will fall with a certain likelihood). For long-range forecasts it is better to be "directionally correct" than "precisely wrong." 5. Accept that your forecasts may not always be popular. A forecast is usually not popular because it is technically sound – it is popular when its conclusion is welcome. Unwelcome results are often rejected or heavily challenged. At some point you will face this situation. Nevertheless, the value of the forecast for your company is not to be popular, but to give the most accurate estimation of the future. Technical and organizational challenges make forecasting a demanding task and the life of the forecaster sometimes difficult. But if you follow the rules presented here, along with those presented in the August issue of Pipeline, you have already started to manage and reduce these challenges effectively. Used together, the 10 guidelines will help you make more accurate forecasts and face fewer challenges implementing them in your organization. Want to learn more on this topic? Please contact:  |