Accurate forecasts are the life-blood of any organisation. They can signal whether an organisation is on track to achieve it’s objectives and they provide an early warning of what needs to change if things are not going as planned. Everybody wants accuracy, but it often seems that searching for the Holy Grail would be an easier task. Michael Coveney, FSN writer, and Author of “The Strategy Gap” looks at ways of managing uncertainty.
An interview with the CFO of a top 100 FTSE company revealed that when it came to forecasting “the more I do this job, the more I realise I’m just as likely to be wrong, as I am right”.
Unpredictability has become a way of life. Today, we realise more than ever, that it is the things in the outside world beyond a company’s control that can have the greatest impact on the performance of an organisation. For example: the credit crunch has reduced the sources of finance, which in turn can hold back customers from investing in new equipment that they would have otherwise purchased; a postal strike can greatly affect the level of service offered and increase delivery costs as companies move to alternate, higher cost carriers; or a threat to oil supplies can trigger a sharp rise in fuel prices that in turn increases the cost of the goods used in production and in transporting finished items. None of which are under an organisation’s control or are easy to predict.
In the past, ‘forecasting’ one year ahead was fairly straightforward as the world just didn’t change that quickly, but with today’s on-line global economy where market changing news can travel to all areas in hours, using past trends that belong to a different era as a way of predicting the future, are no longer reliable.
So how do organisations forecast accurately when it’s almost impossible to predict the future?
The answer lies in why anyone would want to forecast in the first place. Forecasting is not really about predicting a set of numbers, as experience suggests that any set produced, is unlikely to happen even in the most stable of business climates. The real reason why organisation’s forecast is so that they can gauge the future business climate and then adjust their actions to suit their role in that anticipated environment. The main difference in today’s turbulent world, is the need to anticipate a variety of potential situations and to then make sure the organisation can survive/thrive in the event of any of them happening.
To do this effectively, forecasting needs several discrete stages: collecting key driver estimates; analysing driver variability; setting trigger levels; defining best/worst case scenario action plans; and analysis of forecast accuracy.
Key Driver Estimates
Assuming that any forecast is going to involve predictions of future revenue and costs, data needs to be split into fixed (e.g. overheads, salaries) and variable (e.g. materials, energy) items. Variable items should then need be associated with ‘drivers’ – those things that directly affect them. For example, sales volume drives material and direct production costs.
The first stage in forecasting is to fill in the fixed costs and estimate the value of each driver. There should be three values for each driver – what is expected, a pessimistic view and an optimistic view. These views must be supported with reasons for each, and should reflect what could ‘realistically’ happen. As to how far forward you predict depends on the business cycle that occurs in the market place you operate in. For some this may be 3 – 12 months, for others it may be weeks. The important thing is to only go as far forward as you can reasonably predict and only on a few key variables.
Analysis of Driver Variability
Once the driver forecast has been collected, it needs to be analysed as to the total ‘spread’ of results that could result – from pessimistic to optimistic. Management experience of what is going on in the market place needs to be applied to see which drivers are at the most risk of missing the expected result. Driver estimates also need to be compared with the original plan to see what impact that has on things like cash-flow and projected financial statements.
Setting Trigger Levels
The next step is to decide on how far outside of the plan, actual results can deviate before management action needs to take place. Obviously as results come in management will be constantly trying to get the best performance, but there comes a time when things go so far that something significant has to happen. For example, if it becomes obvious that key organisational goals are going to be missed then this triggers the need to re-plan or re-think a particular strategy. These trigger levels can be set against each driver and/or against a derived number such as Gross Profit.
Defining ‘Best/Worst Case Scenario’ Action Plans
With the benefit of a range of values where future performance is thought to lie, the next stage is to overlay a number of potential ‘disaster/opportunities’. For example, a company could visualise three potential ‘major’ issues that could potentially arise that would cause a serious departure from the planned results; such as, bad publicity hitting sales; a product having to be being withdrawn; or the loss of a major customer. For each issue, the associated drivers are adjusted to reflect what could happen and the results that would then arise. These then define a range of action plans that could be invoked to limit the impact on financial results. The purpose here is to plan for the unexpected, and be ready so when it does happen there is no need to panic, and management are prepared.
Analysis of Forecast Accuracy
As actual results come in, the whole process is repeated. Key driver estimates are updated, driver variability is re-analysed and best/worst case scenarios are reviewed. It also gives the chance to see how accurate previous forecasts were. If there is a large deviation then it is important to determine if this is due to an unimagined event, being unaware of a change in the market, or something else.
As data builds up, patterns will begin to emerge about how far forward results can be predicted with accuracy. But the biggest benefit of this type of forecasting is the management dialogue that goes into discussing cause and effect and how organisational activities directly impact strategy.
