9.2.1 Issues to consider when forecasting
When attempting to pick up on trends and to forecast future events, there are several key issues which will affect what method of forecast to use and the accuracy of that forecast. These are:
- Timeframe – how far into the future are you attempting to forecast?
- Pattern of d ata – is there a seasonal aspect to the data or other pattern?
- Cost – is there a budget for collecting and analysing data?
- Accuracy – how accurate does the forecast need to be (+/- 10% or +/- 2%)?
- Comfort level – some decision makers will be uncomfortable with quantitative methods whilst others will be unable to act just on intuitive judgements
- Technology available – is an appropriate analytical program available?
- The use of different forecasting methods
- Jury of Expert Opinion
- Most widely used forecasting method
- Used across the whole range of timeframes
- Advantages are that it is quick, cheap and takes into account intuition that arises from experience
- Disadvantages – very subjective, individuals may have personal agendas
- Sales Force Composites and Customer Expectations
- Used mostly in short and medium term
- Over reliance on these methods builds in bias to forecasts due to salespeople being influenced by current events
- Both advantage and disadvantage is that it is very subjective – hence maybe useful if opinions are required
- Exponential Smoothing and Moving Average
- Useful in the short term
- Smoothes out short term fluctuations to gain a good idea of what the direction is for the period of the average. For example, if the period of the average is 7 days then the forecast is probably not very useful for a period much longer than 7 days.
- Will not take into account trends or seasonality
- Straight Line Projection
- Only useful in the short term
- Will not take into account trends or seasonality
- Regression
- Attempts to find a set of factors which will determine a mathematical formula producing a ‘line of best fit’ for the historical data. This line is then projected into the future to make a forecast.
- Useful only if there are significant amounts of historical data available
- Used for the medium and long term
- Can take into account seasonality factors and other patterns
- Not good if the input ‘factors’ change
- Simulation
- Involves asking ‘what if?’ questions
- Often called Scenario Planning
- Useful in times of extreme uncertainty
- Can be used short, medium or long term but will only be useful as a ‘trigger’ to action, e.g. if something happens then we will take this action. Can be very expensive but software programs are making this cheaper and faster
- Life Cycle Analysis
- Used for product analysis
- Based on the notion that all products will go through a standard lifecycle
- Used by determining where the product is in its lifecycle and then assuming that it will follow the ‘normal’ pattern
- Forecasting needs of management
- Forecasting is used across the business spectrum and across a variety of time horizons.
Table 2 Examples of forecasting across business functions and time
Organisational Unit |
Immediate term |
Short term |
Medium term |
Long term |
Production |
Demand of each product, plant utilisation |
Total demand, employee levels, costs |
Budget allocation, buying equipment and supplies |
New technologies, facility investment |
Finance |
Sales revenue, cash flows |
Inventory levels, short term borrowings |
Budget allocations, cash flows, medium term loans |
Capital expenditure, economic conditions, pension needs |
Activity 4
Answer and analyse your answers to the following questions on forecasting information.
- Select an area within your organisation, or one with which you are familiar.
- What types of information do they need to make decisions?
- How could you forecast future trends in these areas?
- What method would you use to collect this information and why?
Activity 5
- Use management information systems to research and report intelligence required by a transport and distribution business operation.
- Prepare a report on a specific activity using and citing all relevant sources of data.
Back Page