Insight

'Smart' rolling forecasting - how do you do it?

Interest in rolling forecasting is increasing. This is evident from a recent survey by Finext among more than 60 leading companies in the Netherlands. How far ahead are these companies looking? And how do you set up the process of rolling forecasting in a smart way?

How far ahead?

The survey clearly shows that the participating companies use different timelines as forecast horizons. But what determines the choice of these timelines? The horizon is related to the purpose of the forecast. The (rolling) forecast serves as an early warning: bring execution in line with the objectives from the budget and strategy in a timely manner.

The horizon of the forecast will then be at least equal to the reaction time to effect corrective action decisions. In FMCG companies, this horizon may be even shorter than one year. But the horizon can also vary within a single company. For example, Shell will use a much longer forecast horizon for the upstream business for exploration of new oil and gas reserves than in the downstream business with refining activities or retailing gasoline at the pump. Therefore, the horizon chosen is specific to the typology of the business, or business unit.

'SMART' forecasting

Up front, there is no blueprint, no one-size-fits-all process that can be applied everywhere. But there are some working ideas for streamlining forecasting.

We abbreviate these improvement opportunities with the term SMART:

  • Start with Snapshotsof operational plans
  • Measureforecast quality systematically
  • Commit the business on corrective Actions
  • Reducethe amount of detail
  • Ensure a Time-EfficientProcess

Snapshots as a baseline

Operational processes and systems support the business on a daily basis and contain plan data over the forecast horizon. At a defined point in the forecast cycle, a picture or snapshot of this operational plan is taken as the starting point for the financial forecast. This plan data can include your order book and sales plan. The sales plan and the product launch plan are input for the sales projection, taking into account timing, sales volume and sales prices (= "P * Q"). In addition, you take into account special sales prices and discounts for promotions, if this is a material part of the total demand. This snapshot is created financially through a mathematical calculation model as an initial baseline for the balance sheet, P&L or cash flow statement. At that point, operational plans are integrated with the forecast. Deviations from budget, previous forecast and strategy are quantified.

Measure forecast accuracy

A forecast is effective if it provides the most reliable estimate of expected deviations from budget targets. Reliable or accurate means that deviations from actual figures are acceptable. There are always deviations between forecast and realization of KPIs such as revenue or EBIT. In this sense, the forecast is always wrong. A quick win to increase reliability is to eliminate bias, a systematic under or over forecasting relative to your actuals. We recommend measuring and reporting forecast accuracy for transparency into forecast quality.

Direct corrective actions

The forecast must be reliable and inform managers of adjustments if agreed targets are not met. In doing so, you increase the realism of the forecast with a powerful dialogue that leads to actions. Basically, it's about creating mutual understanding of operational plans and forecasts by having an effective dialogue during the forecast review. We call this a Performance Dialogue: the communication is not one-way but two-way. You are open-minded and listen to each other. There is mutual respect and openness about the assumptions on which the budget or forecast is based. Moreover, risks, opportunities, ambitions and feasibility of objectives are shared openly. All this is based as much as possible on facts, but with room for soft opinions. How to set up such a dialogue can be read in the article 5 'must ask' questions in the planning process.

Reduce information

Predictive capability varies over the planning horizon. The level of detail may be higher for early periods than further away in time. The frequency of forecasting is also related to materiality and volatility of forecast variables. Materiality refers to the relative level of a given cost category. Volatility has to do with uncertainty about the development of costs and revenues.

You can differentiate all line items as described in the table below, speeding up the forecast cycle and reducing effort.

Furthermore, you can also differentiate the frequency of the forecast for the short-term and long-term where you update the short-term forecast more frequently than the long-term forecast. For example, we see in practice that companies report a monthly "light" forecast on the qualitative risks and opportunities that materially affect the expected financial results from the previous quarterly forecast.

Time-efficient process

Decouple forecasting and month-end closing to free up more time and focus on forecasting. Forecasting should be completed before the period end. You avoid that the forecast that is partly based on the operational plans for the current period will be adjusted based on the actual figures for one period. The question is whether the deviations over this period are material over the horizon of the forecast. You avoid an iteration that unnecessarily prolongs the lead time of the forecast cycle.

In addition, work in a disciplined manner according to a tight process with a clear calendar. All steps in the forecast cycle are worked out and communicated. Thus, steps that do not provide value to forecast users should be eliminated. In addition, a calendar gives your organization time and attention to assess the impact of current operational plans and corrective actions already underway. That way, you can make timely decisions about changes to operational plans.

From beautiful wish to actual doing

All in all, the survey shows a clear movement among Dutch companies. Where for years there was mostly a lot of talk about rolling forecasting as a nice ideal, a number of companies are now actually deploying it. And given the enthusiasm of the participants at the well-attended event at the RSM, Rotterdam School of Management, where we presented the results, more companies are likely to take the first step soon.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.