Rolling forecasts, or Who ordered that?


I have previously dealt independently with issues of forecasting, planning, and budgeting in separate posts, and the time has now come to pull them all together in one place and just come out and say what I really mean. This integrative post was prompted by a recent invitation I received to address the topic of the benefits of rolling forecasts, where my starting point is always the wise words of David Axson, who said, if you’ve got problems with your forecast process or accuracy, you are not going to fix them by simply multiplying and repeating the same process on a regular, rolling basis. If you work backwards from what it is you want to achieve, very seldom will ‘rolling forecasts’ be the answer. Or, in the words of physicist Issac Rabi (when informed about the discovery of the muon): "Who ordered that?"

First, let’s recap the terminology:

- TARGETS: What you’d like to happen
- FORECAST: What you think will happen
- PLANS: What you intend to do

… and perhaps most importantly,

- BUDGETS: How you are going to allocate your resources.

Because of the now obvious imperative to maintain an agile approach to planning and execution, I am most adamant in my restrictive definition of ‘budget’. Budgets are not to be used for cash planning, targets and incentives, investments, or cost understanding; there are other, better systems for this, and forcing budgets into double, triple or quadruple duty seriously cripples their resource allocation function during volatile times.

With that in mind, I once again reiterate the proper relationship between strategy, planning, forecasts and budgets. Planning is the central component, informed by strategy and forecasts, with the budget(s) as the primary output.

Scenario planning for most-likely, optimistic, pessimistic, best and worst cases, is composed of the following key components:

- The definition and description; of the scenario itself, sufficiently developed so that everyone can recognize it when it occurs and can easily rule out any near misses that it might be confused with.
- The budget; the reallocation of resources or priorities, associated with the scenario.
- The trigger; what measurable quantity must attain to engage the new scenario.
- The implementation plan; is it all or nothing/ all at once, or are there logical stages or steps.
- The incentives; the management bonus schemes and metrics that replace the old regime and will now reinforce the new behaviors.

When it comes to traditional budgeting, we seem to gravitate to one most un-agile technique that I liken to laying 12 months of railroad track at a time, with the only option when conditions change being to issue freezes: hiring freezes, salary freezes, capital freezes and travel freezes, that please no one, not even the CFO who issues them. And how often has your organization gotten to January without a budget in place, with no railroad track at all? On the other hand, we can’t let the organization run around on ATV’s completely unencumbered from fiduciary governance or unaccountable to changing strategic priorities.

The answer is to lay only as much railroad track as each function requires, which will almost never match up precisely with the 12-month Gregorian calendar (in fact it would be a good idea to purposefully avoid 12-month budgets of any kind, even if the project so indicates, make it 11 or 13 months instead). Capital investments and long-term projects will require a budget allocated through to completion, be that six months, 18 months or five years. Programs, initiatives and campaigns, in support of strategic corporate priorities such as quality, customer satisfaction or innovation, need their own budgets so that they are not relegated to operating in cycle time, borrowing resources from traditional departments, and being treated as step children when instead their resource allocation should reflect their strategic importance to the organization (a formalization that Robert Kaplan calls “Stratex” for 'Strategic Expenditure').

When it comes to departmental budgets, scope is still important. Marketing probably needs the ability to commit resources at least a year out for future events and trade shows, whereas sales or production might be kept on a shorter leash more in step with market ups and downs. Neither do sales quotas need to be twelve months in length either – match them instead with the natural sales cycle of the product or service, be that three months or eighteen; it will do wonders to ameliorate one of our biggest self-inflicted wounds – the year end hockey stick.

If your budgeting process is to meet the ‘agility’ criteria, the first thing you want to decouple from the budget are targets and incentives. You cannot be an agile organization, reacting quickly to opportunities and challenges, if in order to implement a revised budget you must also renegotiate the associated bonus schemes at all levels. Targets and incentives are best managed through the use of a Balanced Scorecard, where they can be tied to inter-related key metrics and to external benchmarks such as growth and market share that better reflect the health and performance of the organization than simply an internal, departmental budget.

Finally, it must be recognized that different forecast sources are better suited at providing data for different time horizons than others. Point of sale and ERP systems are most valuable in the shorter daily-weekly-monthly time frame, whereas field sales and departmental expense forecasts can reasonably extend out nine months, and with marketing, industry, government and third party forecasts relevant in the 1-to-2 year horizon.

The point here is that in this context it makes no sense talking about a ‘rolling forecast’. Forecasting is a continuous process that collects, integrates, processes and incorporates into the planning process the most relevant data from each separate forecast source appropriate to its applicable time horizon. Forecasts – plural - as the enterprise seeks to gain as much information, insight and foresight from as many sources and time horizons as seem reasonable. You don’t force an ERP system into provide a rolling 18-month view when backlog only goes out 3-6 months, or when sales can only be accurate out to nine months, and you don’t ask marketing what sales will be next quarter, you ask them instead how product launches, retirements and migrations will affect revenues a year from now. You most certainly do not want to create and encourage an environment where people simply make up the numbers to fulfill corporate’s arbitrary “18-month” rolling forecast mandate. And furthermore, you don’t want to operate in knee-jerk fashion every time there is some change in one particular forecast source, or rely on just one primary forecast input – you integrate them all into one consolidated, weighted forecast view that collectively informs a continuous planning process.

It is the PLANNING process that is rolling and continuous as new information becomes available. An 18-month planning view makes sense - an 18-month forecast from each and every source does not. Scenarios are developed, extended, amended, updated, discarded, and/or modeled as new forecast data warrants and as strategy directs. Resources are then allocated to the functional departments based on the plan on an ‘as-needed’ basis, with some functions requiring longer lengths of bite-sized chunks of railroad track than others depending on the appetite of the resource consumer and upon changing financial, market and/or economic conditions. When you start with better decision making, which includes better resource allocation, as the goal, rolling forecasts do not emerge as the mechanism, but a rolling, continuous scenario planning approach does.


About Author

Leo Sadovy

Marketing Director

Leo Sadovy currently manages the Analytics Thought Leadership Program at SAS, enabling SAS’ thought leaders in being a catalyst for conversation and in sharing a vision and opinions that matter via excellence in storytelling that address our clients’ business issues. Previously at SAS Leo handled marketing for Analytic Business Solutions such as performance management, manufacturing and supply chain. Before joining SAS, he spent seven years as Vice-President of Finance for a North American division of Fujitsu, managing a team focused on commercial operations, alliance partnerships, and strategic planning. Prior to Fujitsu, Leo was with Digital Equipment Corporation for eight years in financial management and sales. He started his management career in laser optics fabrication for Spectra-Physics and later moved into a finance position at the General Dynamics F-16 fighter plant in Fort Worth, Texas. He has a Masters in Analytics, an MBA in Finance, a Bachelor’s in Marketing, and is a SAS Certified Data Scientist and Certified AI and Machine Learning Professional. He and his wife Ellen live in North Carolina with their engineering graduate children, and among his unique life experiences he can count a singing performance at Carnegie Hall.

Back to Top