There have been many methods of financial forecasting used in business over the decades. Seemingly simpler qualitative processes, such as seeking executive opinion from key department heads or using reference class forecasts to look at similar scenarios from the past, have stood SMEs in good stead previously, along with traditional 12-month balance sheets and sales/cash flow forecasts. But for a fuller picture – especially in these uncertain post-Brexit times – are these likely to be deployed alongside more sophisticated quantitative planning methods? Do the likes of driver-based planning or rolling forecasts stack up for SMEs?

Driver-based planning takes the approach of identifying an organisation’s key drivers to create business plans that mathematically model how potential success within the business may be affected by different variables. The goal, according to the What Is encyclopedia, is to focus the plans, in as objective a manner as possible, on criteria most capable of driving success. It says the models may be created with spreadsheets or with more advanced data modelling software applications.

Deloitte, in its web Q&A Driver-based planning: Is it the right approach for your company?, agrees that driver-based models can be set up in low-tech environments using spreadsheets, although it recommends an integrated platform to get the most value from the exercise and “see the consensus forecast across the company, measure performance against drivers and run a distributive process”. Driver-based planning, it suggests, needs alignment across departments in order to reap meaningful insight (see box: An academic question?). As such, driver-based planning works when data is timely, and as Deloitte says with “variance analysis being done after month-end close”. Stale data makes a change of course difficult if the business requires it. It should also be arranged around multiple driver levels.

Deloitte acknowledges that there is no set number of drivers or clear practice that can be applied to everyone, and that driver-based planning is not appropriate for all circumstances. This is especially the case where costs are more likely to be fixed than flex with volume (where traditional or choice-based planning might work better).

Simon Bittlestone, managing director of corporate performance management software company Metapraxis, agrees that the ability to flex/test the plan and shorten the planning cycle is one of the many advantages of a driver-based method: “It also cements the understanding of the business model among the cross-functional team and helps everyone discuss and agree what action to take.”

A survey of CFOs and other senior executives at more than 30 multinational companies was published by business intelligence and corporate performance management software vendor Prevero in January. It revealed the extent to which driver-based planning and rolling forecasting have become an embedded part of corporate monitoring and control.

The survey showed that although all the respondents were using rolling forecasts for sales, driver-based planning was used for almost two thirds (65%) of production operations and more than half (55%) of sales operations, with time savings from shortened budget cycles cited as the primary motivation.

Traditional planning models were criticised for taking a long time to produceand for containing out-of-date assumptions. In contrast, driver-based models focus on a small number of key drivers and therefore can produce a forecast or plan much more quickly, making them ideal for ad hoc analysis of multiple scenarios.

Len Jones, group FD at Practical Car and Van Rental group, suggests that this kind of statistical-based forecasting is very useful in times of economic uncertainty. He says FDs have been instrumental in trying to align business costs in proportion to output, so that any reduction in revenue is reflected in the appropriate costs so that margins are maintained.

“Focusing on the key drivers helps realign management to key priorities and sources for growth,” says Jones. “But an FD’s perspective can be different from that of a CEO – sales per salesperson and performance-related criteria may be what is required, rather than sales ratio.”

Focusing on drivers for planning purposes turns the firm into a wealth creator rather than a cost administrator. Companies should not be too concerned about budgets per se, but rather the sensitivities of the cost structure to changes in business drivers – the “what if” scenarios, as Jones describes them.

“At SME level, senior managers/owners know the business instinctively; they don’t need a driver based forecast to tell them the firm is heading in the wrong direction,” he adds. “In a larger business there is still this instinct, but the driver-based plans are used as a method of controlling behaviour at individual or departmental level.”

According to Bittlestone, the principle of simplicity applies to the implementation of rolling forecasting. Forecasting a full 12 months each month can be impractical if it relies on a time-consuming manual, zero-base process. Much of the advantage of an up-to-date forecast is lost if it takes weeks to create.

While there is still plenty of enthusiasm for rolling forecasts (certainly given the ability to compare iterative results with static data from previous periods) the perception that they are time-consuming persists. Bittlestone adds that forecasting horizons might need to be adjusted for different parts of the same business and that beyond the horizon at which accuracy is no longer possible, it can be more useful to calculate and use a risk range.

Stephen Pugh, FD at Suffolk-based brewer, hotelier and wine merchant Adnams, says his company has debated but not implemented rolling forecasting: “We still do an annual budget and quarterly reforecasts, which work well for statutory reporting but have less relevance in management reporting where a year is obviously an arbitrary time period. The issue with reforecasting is that businesses with uncertain or lumpy sales may need constant liaison with sales staff, which is a distraction and can lose goodwill.”

According to Pugh, businesses with large numbers of small sales may be more predictable and can be effectively forecast by the finance team.

Gemma Davie, senior director of financial planning & analysis at CA Technologies, says: “Rolling forecasts force management to look at future issues and take action early, particularly when combined with driver-based planning,” she explains. “They are typically most effective when focusing on 10-15 business drivers as they allow for changes in business conditions. But in the short-to-mediumterm they can be time consuming as traditional methods will still be required – particularly the budget-setting process – and this is one of the main reasons for failed implementations.”

She observes that since driver-based planning is based on historical trends, it could produce inaccurate forecasts if those drivers change. It also does not consider unforeseen external events.

The challenge is to find the 20% of business drivers that explain 80% of outcomes, adds Larysa Melnychuk, founder and managing director of London FP&A (Financial Planning & Analysis) Club. “The larger the company the less likely it is to have a full driver-based planning model,” she says. “Reasons for this include the existence of multiple legacy systems, lack of consistency between business units and the sheer number of people working on financial planning.”

Melnychuk suggests that about 20% of companies abandon rolling forecasting after initial implementation because it does not add value: “This happens when it is not undertaken through key business drivers or when the programme is not specifically designed for the company.”

Bittlestone believes implementations fail because they are over-complicated and poorly communicated by finance teams seeking perfection: “The model should be high level, top-down and simple. It is not just about mapping the drivers and presenting it to the business, it is a process that the business must be involved in so everyone agrees and understands it. Without this, engagement in the subsequent planning exercise will be low and results poor.”

Data and information are critical to driver-based planning implementations – but they are often in short supply and/or buried in legacy systems and the skills and time to make full use of what is there are limited, warns Develin Consulting director, Paul Clarke. This makes timely forecasting challenging unless there is a significant investment in better systems and quicker access to data.

He adds that accounting systems are often not well aligned to driver-based planning and rolling forecasts: “Plans have to be translatable into basic accounting controls (eg, the impact on cash), but they must also allow measures to be performed on a consistent basis – such as the unit cost to serve a customer and how that cost might differ if an ‘unwanted’ driver of cost pops up, such as a customer return.”

Clarke also suggests that some businesses attempt to introduce driver-based planning and forecasting while still hanging on to the traditional way of putting the budget together.

Jones says most traditional budgets start creaking after the second quarter. “Appropriate time horizons are KPI or driver-dependent. What is required is a trend analysis and a time horizon that says there is enough data for the predictive modelling to be supported as accurate and robust or require changing.”

Davie notes that if a company is listed, there will be rules and regulations relating to how often it will need to communicate forecasts to the market, such as earnings calls (including forwardlooking statements). “Availability of information will also affect how far into the future an organisation can plan – if less information is available, long planning horizons would not be recommended,” she concludes.

Paul Goodwin, professor of managerial science at the University of Bath, agrees that driver-based planning has some limitations. Although providing an evidence-based formula linking success metrics to specific factors enables an accurate understanding of which factors are most critical to success, there is bound to be some subjectivity in the process. For example, the selection of drivers that may have a significant association with success will be based, at least in part, on the analyst’s judgement. Similarly, the form of the model and the amount of historic data that it is fitted to will be based on judgement calls.

Collinearity between the drivers (where two or more drivers are themselves correlated so their effects can’t be easily separated) and relationships that change over time may also pose problems for the analyst. Finally, there may be a resistance to the use of mathematics, particularly where the model conflicts with long-held beliefs.

“The clear advantage of rolling forecasts is that they are updated as new information becomes available,” says Goodwin. “However, when the updates lead to frequent changes to the forecasts this may reduce their credibility in the eyes of users. Frequent changes may also cause some problems for people whose plans have to be revised each time the forecast changes. When judgement is involved in the forecasting process there is a danger that people may overreact to recent ‘patterns’ that are merely a manifestation of randomness.”

The appropriate forecasting horizon depends on the nature of the decision, he says: “Ideally, forecasts should be expressed as probability distributions, but this is rarely done and there is an over-reliance on point (single number) forecasts. In the very long term the level of uncertainty will be such that organisations will be better advised to replace forecasting with methods like scenario planning.”

Download pdf article:

• Light the way ahead
  Finance & Management Magazine, Issue 245, July/August 2016

The ICAEW Library & Information Service provides full text access to leading business, finance and management journals and a selection of key business and reference eBooks. Further reading on rolling forecasting is available through the resources below.