Risk (or uncertainty) is inherent in any business or decision situation in which forecasting of one form or another is undertaken (e.g. sales forecasts, cost budgets, or cash flow). Nevertheless, even though much effort is often invested in building financial models to support such analysis, the explicit modelling of risk is often overlooked (with the analysis being “precisely wrong, rather than approximately right).
The benefits of risk modelling are many. Whilst they are described in detail in Chapter 4 of my book Business Risk and Simulation Modelling in Practice (see here for more), a few key points include:
- Finding out where a base case fits within the full possible range of outcomes
- Building more realistic and accurate models (including event risks and the effect of non-linear model logic, real options, valuing flexibility etc.)
- Using the range of outcomes to inform decision-making linked to risk-preferences, such as the setting of project targets and contingency budgets
- Capturing partial dependencies (e.g. correlated sampling) between uncertain processes.
Whilst many aspects of business risk modelling are similar to those of traditional (static) business or financial modelling, there are also some challenges (once again, these are covered in detail in the book), including:
- Many clients find that the mapping of the nature of the risk/uncertainty into a quantitative model can be challenging. For example, one simple but common pitfall is to treat uncertainties within a probability-impact framework, whereas such a framework is really only suitable for event- or operational-type risks. Another main pitfall is to try to build a full risk model based on a set of risks that has been derived from a qualitative risk register, which is usually inappropriate (as described in the book).
- The organisational context in which the results risk modelling is to be applied often poses a challenge; there is typically scepticism, and/or insufficient (broad) understanding of the meaning and implications of the analysis, so that the interpretation and communication of risk modelling results is often a key concern.
Michael’s consulting work in this area revolves around the building of risk (uncertainty) models, although such activities invariably need to extend beyond pure modelling, typically involving stages such as:
- Initiation and Identification. Using one-to-one meeting and groups workshops to discuss objectives and approach, to engage appropriate staff, build consensus, and to kick-off the work (e.g. to identify key risk drivers, data sources, and conduct detailed project planning and co-ordination).
- Modelling and Analysis. This involves the building of an appropriate quantitative risk model to support the decision situation, to analyse the outputs and determine implications, and to start the communication of the results. This is typically a multi-stage step with interim review points, meeting and workshops.
- Planning and Hand-Over. This involves the development of action plans, the appropriate communication of results, and further workshops and training as necessary.
His training courses in this area are very hands-on and customised, and focussed on specific applications of interest to the client. Participants design and build risk models that are relevant to their own situations, so that the learning is “by osmosis”; one assimilates the relevant concepts in modelling and statistics almost without realising it.
Although much can be achieved without the use of external commercial add-ins (i.e. by using only Excel, VBA/macros, or self-made add-ins built with VBA), very often it is more efficient and more transparent to implement the simulation component of the process by using a commercial add-in, such as @RISK®. This applies to both training situations, and to consulting engagements. Michael is Europe’s most experienced instructor and user of @RISK, having taught over 2000 people in more than 300 seminars since 2004, as well as having done many consulting engagements in that use the software. (He can also user other similar add-ins e.g. ModelRisk®, RiskSolver® or CrystalBall®).
Contact us to find out more.