In the face of rapidly changing, unstable market conditions, traditional annual budgeting and forecasting processes have become outdated. Making rigid assumptions months ahead of time can often lead to worthless forecasts, as things can turn upside down so quickly.
Being able to reforecast and adjust predictions based on the newest data should be a priority for every organization – which is why more and more FP&A teams are turning to a rolling forecast model.
Rolling forecasts are reports that use historical data to predict future performance continuously over a period of time.
The word `
continuously` is the key idea here. Rolling forecasts are usually 12-month (or sometimes longer) windows of performance. That means that, as a new month is added to the forecast, the oldest month is removed from the calculation. Unlike traditional forecasting, rolling forecasts go beyond the confines of one financial year.
That makes them inherently more forward-looking than traditional methods of forecasting. A rolling forecast approach can play a central role in positioning a company for sustainable growth. It allows companies to make adjustments on the fly and react to changing market conditions with agility. But these forward-thinking analyses can quickly become unreliable as a result of a number of missteps in an organization’s approach to them.
In general, the root cause of the problem lies in the disconnect between the finance function and other parts of the business. This can lead to a number of issues affecting the quality of your forecast.
Today we take a look at the four most common challenges associated with rolling forecasts and, more importantly, how you can address them.
Many organizations think they’re creating rolling, dynamic forecasts when they update actuals throughout the year, but if they aren’t also pushing out the end date, they aren’t really informing the business about the long-term consequences of current business decisions.
This is often called “forecasting to the wall.” For example, many organizations create a 12 month budget and make subsequent updated forecasts as new data comes in, but the time horizon remains the same.
This means that the 12-month forecast you started with in January, will only be a 3-month forecast by the time you get to October. As a result, your forecast will be of limited benefit when making decisions for the future, which is a key aspect of having a forecast.
A better approach is to establish a rolling forecast. Rolling forecasts are based on dynamically adding an additional forecast period every time we update for an additional period of actual results.
For example, if you want a 12 month rolling forecast, when you update January to actual results, you would add a forecast for January of the following year. This ensures that you’re forecasting for the same amount of time and not hitting the metaphorical ‘wall’.
One of the most effective ways to improve forecasting accuracy is to give your entire organization access to all the relevant data in one place that integrates with all of your tools.
By doing that, departments such as finance, sales, marketing, and other business operations can track how different drivers are affecting each other and course-correct as necessary.
The more integrated your data is, the more the accuracy of your forecast increases. Unfortunately, many organizations work with multiple siloed sources of information that often provide conflicting numbers.
In most cases, they are relying on people manually copying data from one tool to the other or digging through spreadsheets to find what they need. Sometimes, there are data discrepancies in these tools, forcing your team to spend precious time checking the validity of the data.
Direct data integrations can reduce the need to referee departmental arguments over data validity and help you get buy-in from the whole company for your forecasts.
From a forecasting perspective, your organization will have a much easier time making changes on the fly if you’re working with real-time data.
Similar to the previous point, if you’re relying on Excel and Google Sheets and manually copying information from one place to another, that’s bound to take up much of your team’s time. It’s also much more likely to lead to errors.
And by the time you collect everything you need from spreadsheets, it’s often too late to make an accurate forecast.
The point of rolling forecasts is to improve predictions based on the newest data available. If manual processes are preventing your team from collecting and organizing information on time, you’ll be missing out on their biggest value.
Instead of focusing on searching for patterns, understanding the relationships between your key business drives, and finding insights, your FP&A team will be stuck preparing everything you need for a high-quality forecast.
When preparing rolling forecasts, it is critical to find the right balance between accuracy and the time required to update them.
If you develop a very detailed forecasting model that takes too long to update, you’re unlikely to have the resources to continue updating it. If you can’t keep your forecast current, then it is no longer able to provide insight for decision-making.
The 80:20 rule often works well in determining the level of detail needed in a forecast model. You can get 80% of the benefits of the perfect model, by putting in 20% of the time required to build / maintain the perfect model.
Automating as much of your forecasting process as possible through integrations can significantly reduce the time taken to update forecasts.
While you don’t want to overbuild a model, as mentioned above, you do want to assess it over time and look for ways to improve it. So look back and compare your historical forecasts with actual performance.
There are many reasons you might find variances between actual and forecast results, including:
Variance analysis can help you gain a better overview of your business’ financial performance and forecasting accuracy.
It helps you see patterns in the relationships between key drivers in your business, like the number of new leads generated by marketing, customer satisfaction rate, sales expenditures, and others.
It also allows you to understand how these relationships have evolved over time, e.g. how your customer retention rate has changed in the past couple of years relative to another driver and highlight why that’s the case.
If your assessment of historical forecasting accuracy highlights model errors, or drivers that are not accurately modeled, these can be updated, leaving you with a model that is better able to support decision making.
To address the mistakes mentioned in this article and improve your rolling forecast process, you should aim to:
Ed note: Want to dive deeper into essential FP&A topics? Check out OnPlan’s most popular posts of all time here:
In this article, we’ll cover some board report essentials: their purpose, general structure, and writing…
The story of how OnPlan was founded, our CEO playing basketball with P Diddy, and…
Congratulations! You’ve raised your Series A funding and now it’s time to think about how…
Charlie Liu is the Director of Solutions at OnPlan. His responsibilities include architecting Apps for…
Learn about SaaS profit margins and see industry benchmarks to find out more about your…