12-month rolling forecast

[benito]

I am struggling on the 12-months rolling forecast which might need some help.

1. Unlike an annual budget, a 12-months rolling forecast requires continuous update on a monthly basis. It's January now so the 12-months rolling forecast will be February 2025 to January 2026. In February 2025, we will have forecast figures for March 2025 to February 2026. But do we usually need to revise all forecast figures from March 2025 to January 2026 that we did in our January 2026 version, or simply adding February 2026 as the last period in the forecast is suffice? If we need to revise the forecast for all periods that we did previously, it will look like we have to perform an annual budget each month and it will definitely be a disaster. On the other hand, if we just need to add a column and plug in the figures for the 12th month, the figures that we prepared half a year ago on the rolling forecast may lose their accuracies (which is similar to a budget). For example, the Oct 2025 forecast that we prepared now will lose its accuracies when we are in Sep 2025. So what is the best practice of preparing the 12-month rolling forecast?
2. It's very easy to use the "forecast" formula from excel to do a forecast. However, when we need to perform vertical analysis between Actual vs. Forecast, we will have no idea about the variances for each period because we have no idea how excel calculates the forecast for each period. So, do we usually need to perform the forecast by digging into each line items including the cost drivers (similar to budget)? Again, if we need to keep revising the rolling forecast for each period, it will be a super headache each month.

Just wonder what the usual practice is for preparing the rolling forecast and if there is an easy way to work on it. Thank you.

 

[DrStrangeLove]

Whether you need to re-forecast all future months in the forecast depends on how much the difference between the first actual month vs. the first forecast month would ripple forward. For example, if your actual cash burn is a lot higher than forecasted, it causes a cash strain in the next month. So cash, working capital and inventory (I don't know your industry) would need to be re-forecasted. Longer-term, discretionary items like capital expenditures or sales of fully-depreciated fixed assets probably won't need to be re-forecasted. Figure out a robust, easy to update design for your forecast process, and find a rhythm for updating the forecast.

Your actual/expected analysis will need to dig into material items where the actual/expected variance is substantial. ("Material" and "substantial" are a matter of judgment, so YMMV.) But yes, you'll need to compare the variance to the item's drivers and flex the expected for the actual drivers. Agreed, it's a pain. But it's the right way to do it. Design your forecast system to show the right level of detail and to be easy to roll up, and you'll save yourself a bunch of pain each month.

There's a reason lots of companies look for people who are good at FP&A.