It’s worth mentioning at the outset that a forecast is a short-term view of what is most likely to happen, irrespective of what targets have been set. It takes into account the resources available as depicted in a budget, and sees the business environment in an assumed state.
Forecasts and Accuracy
The accuracy of forecasts depends on a number of factors, one of which is luck. Just because we have a good feel for the market and our role in it, doesn’t mean to say we can actually predict what will happen. At best we can get a good idea of what may occur, but even then the forecast could be completely wrong as future performance will depend on factors that are both unknowable uncontrollable. For example, the impact of the weather, comments made on social media, or the actions of a competitor can all greatly impact an organisation’s performance – but that doesn’t mean to say we shouldn’t forecast.
What forecasts do is to give an indication of the most likely outcome, within an assumed business environment. For this reason, they should always be a range of values as this isn’t an exact science. However, what is important is that forecasts should be the subject of reason and backed up with some form of empirical evidence.
With this in mind, forecasts can be greatly improved by:
- Involving more people who are relevant in the process.
- Challenging assumptions on what drives performance.
- Collecting the detail behind the numbers (workload, resources and outcomes) to give management a more comprehensive story.
- Producing trends to see if submitted forecasts are logical.
- Allowing managers to determine the variability of individual forecasts.
To see how this works in practice, the following sections take each point in turn and consider the role played by technology.
Interestingly, for most organisations, that technology is a spreadsheet. This probably reflects various reasons: as everyone has one there is no incremental cost, and they are fairly easy to set up at the outset. However, they do have limitations, particularly in the areas of managing and analysing submissions.
Assuming that something more specialised has been selected, technology can play a key role in ensuring forecasts are realistic, although that does rely on it being used in the right way. To take each point mentioned above in turn.
Allowing more, relevant, people to be involved:
Forecasts, unlike targets, are a bottom-up process. It requires those closest to business activities to offer the most realistic view they can on what is going to happen. To do this requires a system that is multi-user, and that ‘knows’ which users are responsible for what departments along with the associated measures. Most planning systems have this capability ‘built-in’.
Administrators first define what they would like to capture, and from which departments. The level of detail can vary according to department; for example sales departments may be required to enter sales and expenses; the IT department would need to enter just IT related costs, while human resources (HR) might provide salary information for the complete organisation.
The workflow capability then ‘hands out’ these requests and automatically provides a tailored data entry screen to capture what was requested. The security system automatically filters out data they are not allowed to see and can provide comparatives – for example the budget to help with submissions, which can be viewed and not changed.
Once a user has entered the data, he/she can submit it for approval and at the same time will be locked out from making changes. Because all data is held centrally, administrators will always know the status of any submission and can perform a consolidation at any time to see the latest position.
This level of automation means more people can get involved, but without much effort being required from those overseeing the forecast process.
Challenging assumptions on performance drivers:
When collecting forecasts, costs and revenues can be typically split into two types:
- Those that are relatively fixed, such as rent, rates, loan repayments and annual maintenance revenues.
- Those that are variable and dependent on other factors. For example, raw material purchases and some manufacturing costs are often dependent on the volume of products made, which itself could depend on the volume of product ordered.
Fixed amounts are set months and maybe years in advance, and these can be entered into the forecast system when known. Although variable measures cannot be predicted in this way, they can be modelled. Each organisation achieves its aims through a series of connected business processes. For example, sales revenue is the product of marketing, lead generation, sales calls, customer references and signing contracts. Similarly, production costs are the product of buying raw materials, fabricating them into products, packing and shipping them to customers.
Sample business process for sales showing linkage between workload, resources and outcomes.
Each activity within a business process is linked – the outcome of one activity becomes the input to the next activity. Supporting these business processes are a range of support activities, such as IT, HR and finance. The efficiency/success of an organisation depends on how its business processes connect inputs with one activity to the output of another.
In measuring business activity, the figure below illustrates the connection between workload, resources and outcomes:
Workload such as publishing adverts or attending sales calls consumes resources, although it could be argued that the level of resources dictates the work that is conducted. The net result is outcomes that are generated. Most organisations have the following four key business processes, each with an ultimate outcome that contributes directly to business goals:
- For the sales process this is typically revenue.
- For production this could be the total number of orders fulfilled.
- For marketing this could be market share.
- For product development it might be the number of new products/services in the pipeline.
The linkages can be modelled, so that entering information such as workload or targets to be achieved are then used to generate the resources required. This is sometimes known as ‘driver-based’ planning. The models allow management to assess historic performance as to what drives success, and to challenge the way business processes are conducted and resourced in the future.
From a forecasting viewpoint they can be used to determine whether the level of resources being applied is likely to achieve a particular outcome; or if the outcome being forecast is reasonable based on the resources and workload assigned.
Collecting the detail behind the numbers:
The third capability a forecast system can provide is in giving management a more comprehensive story behind a forecast number. It’s simple, for example, to enter a forecast of sales revenue, but the key question is the likelihood that the figure entered will actually be achieved? Quite often forecasts are entered under pressure; for example if the forecast is under budget, then the fear is that senior management will become irate. Yet if the number is above the budget, then there is a fear that targets will be increased in future months. For this reason, some users feel that it’s safer to enter the budget as the forecast.
Putting political pressure aside, what senior managers really want to know is what is most likely to happen. This allows them to be better prepared should the forecast turn out to be correct. The only way to assess reality is to look at the detail behind the numbers. For sales forecasts this may include collecting prospect situations such as the date of the last visit; whether a proposal has been sent; the likely chance of success; and details of the competition. Similarly on expenses, knowing details such as planned marketing campaigns and customer visits can help management assess whether the level of activity being supported is ‘reasonable’ and likely to lead to the outcomes forecast.
To do this, the forecast system must be able to handle user-defined levels of detail and should provide support for text, dates and notes that can be directly attached to individual forecasts.
A sure sign of an unrealistic forecast is when there is a step-change in performance. During the year there may be a steady increase/decline and then in the space of one period, the figure being forecast appears to buck the trend. There may be a good reason for this – for example the launch of a new product or some other event – but there must be a reason. This sudden change in direction is often seen in budgets, where the first period target is much greater than the previous period actual or where sales targets show a ‘hockey stick’ effect towards the year-end.
There are various ways in which a forecasting system can be set up to challenge these types of submissions:
- As forecasts are entered, the system can place them on a timeline that shows them as a continuous chart following on from actual results. Where the forecast appears to change direction, then the system should prompt the user for an explanation.
- As in the previous point, forecasts can be superimposed on a chart showing last year’s actual results. The aim is to highlight any trend not conforming to past experience, so that the user can give a reason as to why the forecast is different.
- In the case of one organisation, the forecast system performed statistical analysis of the previous two years of actual results. This was then used to provide the user with a measure that showed the statistical degree of confidence on any forecast entered. For example, it would tell the user who entered a sales forecast the percentage likelihood of the forecast being achieved, based on statistical analysis. It was found that this type of information made users think, as they knew that senior managers would be getting the same statistical data. The result was far more accurate forecasts than when this information was not provided.
- As mentioned earlier, the forecast system can request and plot dependent measures together. For example, when requesting the forecast for an outcome, the dependent resource and workload measures should be displayed. Given that some outcomes are delayed in time, the system would need to show measures for a number of periods so that their effect can be assessed.
Individuals who oversee the forecasting process should also be provided with all the above capabilities. This could be in the form of alerts, charts or more formal reports. Analysing and approving forecasts is just as an important role as submitting them.
Assessing variability of individual forecasts:
The last area where a forecasting system can help is in keeping a history of prior forecasts by person/department and measure. As actual results come in, the system can contrast these with previous forecasts to determine how far into the future a particular measure/department can be determined with a certain level of accuracy.
Forecasts that are inaccurate although made just a few periods ago, or those that fluctuate widely with each period, could reflect a number of reasons. For example it could indicate that the people involved don’t really know what’s going on. It might be due to political pressure, with no-one willing to be the bearer of bad news. It could also mean that some measures cannot be predicted that far into the future, being dependent on factors outside of the organisation’s control
Whatever the reason, knowing that a number cannot be forecast is just as important as knowing what the number is going to be. It signals that contingency plans are required to offset any performance that could jeopardise overall results.
This article has attempted to outline the benefits of a forecasting system, but it is only as good as the way in which it is set up, the truth of the data being entered, and the analysis that is applied.
As to whether which forecasting system is best, then that is dependent on a number of factors outside the scope of this piece.
Coffee is the second most traded commodity in the world and is subject to plenty of price volatility. Costa Coffee's treasury manager discusses how the company hedges pricing and foreign exchange risks to minimise volatility.
Ireland is a potential beneficiary as businesses relocate from the UK post-Brexit, but this is undermined by the threatened disruption and complexities that increasingly appear to be a likely consequence.
A US study, based on the quick service restaurant chain Chick-fil-A, offers conflicting evidence on whether a TMS is the best option when upgrading from Excel-based forecasting.
With four hikes since late 2015 - three of them in the past six months - US interest rates are moving higher in line with a strengthening economy. However, for many American corporates the trend has raised concerns over their working capital management.