GlobalBX

What is Forecasting?

A definition of Forecasting: – A strategic planning approach that supports management’s attempts to prepare for near and distant future events, based largely on past and present year’s data and trend analysis. Forecasting starts out from compiling a set of assumptions based on the experience, judgment, and knowledge of the management team. These estimates are then expanded over the coming years and more immediate months using one or more of a variety of techniques.

Basic forms of forecasting are part of our day-to-day life, we check the weather forecast before we go on holiday and the tide table before we go down to the beach. A small business might find it necessary to understand the seasonality of its industry or when it reaches the point where it out-grows the serviced offices it occupies and needs to start looking to rent bigger premises.

Why is Forecasting important for Businesses?

Financial forecasting is important for several reasons. First, it enables management to change course at the right time in order to reap the greatest benefit. It also helps the company prevent losses by making the proper decisions based on relevant information. Organizations that can create high quality and accurate forecasts are able to “see what interventions are required to meet their business performance targets”

Forecasting is equally important when developing new products or introducing new product lines. It helps management decide whether the product or service will be successful. It is a vital tool for the company to prevent it from spending time and money developing, manufacturing, and marketing a product that may ultimately fail.

Demand forecasting is an important tool of performance management as other management processes like the budgeting process, procurement and HR management relate to it and build on it. An accurate forecast will result in more accurate budgets which in turn have been found to be positively related to performance.

Before getting involved and investing substantial amounts of time and resources in an attempt to predict the future, it is important for management to get a good understanding of the company’s trading environment and the forces that work on the company from the outside and have a direct effect on its performance. Useful tools for analysis are for example a PESTEL analysis and the model of Porter’s Five Forces both of which can be used to map the company’s environment.

Trends in demand are usually influenced not only by decisions made by the company and actions resulting from those decisions but quite often come as signals from the outside. In many cases, the company has absolutely no influence on these potential key drivers. Some months ago a high court ruled that car insurances must not differentiate between male or female when setting the insurance quote based on accusations of sexism which subsequently destroyed the business model of Sheila’s Wheels – a car insurance company specialised on cheaper insurance for women which was based on empirical evidence that women have less accidents than men. Whilst Sheila’s Wheels may have seen this coming in the months leading up to the verdict, they would not have been able to predict this when they set up their business plan.

But there are changes to the environment which can be predicted in forecasts and often it is the forecasting process which gives the manager a better understanding of the greater picture.

Different Approaches to Forecasting Demand

There are two main approaches to forecasting, qualitative methods based on opinions, experiences and even best guesses and quantitative forecasting techniques which use historical data to model forecasts. Neither of these approaches is proven to produce the more accurate results, it depends very much on the context and quite often a combination of quantitative and qualitative techniques can produce the most accurate results.

A qualitative approach involves collecting and appraising judgements, individual or group opinions and even ‘guesstimates’. These collaborative methods usually involve panels of experts and there are various formats, the most highly regarded one being the Delphi method. Like the conventional panel approach, the Delphi method involves a number of experts in the field under review. But in order to eliminate the effects a group discussion can have on the result (for example the bandwagon effect), participants are sent a questionnaire to fill in. The questionnaires are returned and analysed and the collected, summarised information is then sent back to the panellists with the request to review their own answers in light of the information now available. This process is repeated a few times until a consensus is reached or at least the possible results have been reduced to a narrow selection. This is a time consuming method which also requires a lot of resources – not an approach often used in a business context but much rather in politics for example or in focus groups when doing market research.

Benefits of collaborative forecasting are that a consensus between different individuals and potentially different departments is reached which supports communication and ensures that everyone involved works from the same assumptions and plans. Complex issues are overcome by cooperative solution seeking and historical data is complemented by in-depth expert knowledge of present operations.

Issues around collaborative forecasting can form around individual personalities or relationships, some panel members being more extrovert, influential and louder than others and this influencing the outcome of the discussion. These panel members being followed by their members of the group who develop a tendency to agree with the informal leader is the effect known as the bandwagon effect. Bigger groups often take longer times to agree and find a consensus so the accuracy of the outcome is potentially already flawed by compromises made during the process. Another issue with using people from within the processes or the departments involved is that the individual’s underlying resistance to change and an increased work load may impair their judgement.

Qualitative approaches involve the analysis of historical hard data; there are two main approaches to qualitative forecasting which are time series analysis and causal modelling techniques. The former approach involves a number of techniques which plot a variable over time, try to identify reasons for underlying variations and after removing those attempts to forecast the future as precisely as possible. Underlying trends such as growth or seasonality have to be identified and removed before more basic forecasting techniques such as Moving Averages can be successfully applied and actually produce some viable data. At the more sophisticated end of the spectrum of techniques we have Winter’s Exponential Smoothing technique which includes three equations each with a smoothing parameter with the aim to remove the effects of patterns like level, trend and seasonality. Whilst this technique can produce pretty accurate forecasts, the calculations involved are complex and therefore not particularly suitable for SME’s.

Causal models look at the relationships between a variety of variables and aims to establish how one influences the other. A relatively simple causal model would for example examine the relationship between the surge in demand for wellington boots in summer (which at first may seem unusual) and the annual music festival season (think Glastonbury etc.). But in real life, causal models are usually much more complex involving a large number of factors and relationships to analyse and are therefore only rarely suitable for small businesses.

So in essence, forecasting can be as simple and basic or detailed and sophisticated as required, there are many very different approaches all aiming at the same goal – to predict the future demand or outcome of a given situation or scenario as part of the planning and management process and to pro-actively manage the performance of the company.