Driver-Based Planning: Review the model
Posted by Michael Coveney
In this series of blogs I am providing a comprehensive approach to the design, construction, roll-out and maintenance of a modern driver-based planning solution. In this last blog I look at the typical life-cycle of a model.
As mentioned in my last blog, rolling out a planning solution should be done in stages. This will enable the model to be established and checked for usability, as any problems if not solved quickly, will cause the model to be rejected. A suggested development life-cycle would be:
- An initial, basic driver-based model, where driver values can be entered by FP&A to predict outcomes
- Model cascaded down through the organisation to be used at a departmental level. Actual data is automatically loaded and forecasts consolidated.
- More sophisticated model generates driver values and/or challenges users on the value of drivers entered. This may require additional, linked models.
- Model forecast timescales reduced (e.g. daily forecasts) through automation
- Model extended to cater for departmental specific requirements
As can be seen, planning models are rarely static. They start out simply but as reality and experience develop, models become increasingly complex, as anyone who has developed a spreadsheet-based model can tell you. After a while changes will be required either because something is missing or because the world in which the model was created has changed. Those changes can be categorised as:
Internal to the organisation. This could include new users, departmental structures, and introduction of new products/services, which need to be reflected in the model
Changes within the market place. This could include the impact of competitors and/or market expectations which may lead to the change or introduction of new business drivers.
Change in assumptions. Models are created with imperfect knowledge and so as they are used and results obtained, they will need to be adjusted in line with our better understanding of the organisation’s business model and the way in which customers perceive the value we bring.
As those changes are addressed, models can start to lose their focus with the inevitability that at some stage they will need to be re-written if they are to retain their value. This should be planned into the life of the model. In effect their development is never finished.
A good way to approach this is to form a team of stakeholders who meet regularly to discuss the performance of the model and who direct its development. That team should cover the key stakeholders involved in using the model. The questions they should discuss would include:
- How far into the future does the model produce ‘accurate’ forecasts? Is there any reason why this cannot be improved?
- Does the model reflect the way in which the business operates? Are the internal drivers the correct ones?
- Does the model reflect the current and future market place? Are the external drivers the correct ones?
- What could be done to improve usability?
- What could be done to make the process faster and more efficient?
The answers to these questions should be documented and used to create an updated development plan, which can then be shared among all users.
Planning systems are at the heart of an organisation. Through them, they can speed up the forecast and decision-making process.
They can reduce the politics and guesswork out of plans by basing them on well-defined (and hopefully agreed) assumptions on how the organisation operates in the current business climate.
With such an important task, they need to be properly resourced and maintained.