Joanne is the CFO of a private equity support appliance manufacturer. For Joanne to do her job effectively, she needs to have a (relatively accurate) measure of how much the company sells by the end of the year.
In that prediction, Joan usually relies on historical trends. There's no more. This year, AMID tariff whiplash, lower customer sentiment, higher interest rates, and the fear of a recession – historical trends have proven somewhat meaningless.
And that's a big problem for Joan and almost every other CFO.
Without a precise way to predict revenues in a volatile economy, too many CFOs will catch you off guard by a sudden cash flow shortage, or a company's killer. Some people may have to touch the dreadD “PE Third Rail” By soaking in a revolver.
How can Joan predict his revenue without looking at the rearview mirror?
It's a collaboration.
Yes, finance must own the process, but that's not a solo act. Accurate revenue forecasts collect inputs from across different features to form a comprehensive view of your business.
For example, sales must own the hygiene and governance of pipeline data, monitor pipeline health metrics, and understand customer forecasts and sentiment. This provides this information on a scheduled basis for input to predictions. The operational team must be responsible for predicting supply-side visibility. As tariff disruptions continue to affect supply chains, we need to understand data on backlog visibility and fulfillment constraints.
However, CFOs are not just for aggregating and entering data. They're there to analyze it: evaluate the inputs and pressure test to make sure the predictions they're generating are realistic and accurate and defensible enough to guide the company's strategy.
What is the scenario?
Hip hop group A Tribe Called Quest knew one or two things about predictions. The scenario, or rather the scenario plan, should be burned into a forecast process.
This is especially true in markets where tariff whip and fear of recession are dominated. To achieve reliable forecasts in unstable markets, CFOs need to explain factors such as pricing changes, demand delays, and supply chain disruptions. You should also conduct and include risk and opportunity analysis that draws pictures of what happens if things get better than expected or worse yet.
And this exercise should not be a slide of Joanne's “Monthly Business Review” material. Teams need to consistently track these insights by delving deeper into customer behavior and coping with dynamics and pricing patterns. The R&O analysis then forms the basis for various financial results and informs gross profit, EBITDA and cash forecasts.
Rely on new releases.
The forecasting process should occur twice or more times a year. Real-time liquidity visibility is particularly important in unstable environments, meaning that forecasts often rely on numbers that do not accurately reflect current business reality. CFOs should take a more dynamic approach that relies on 12 months of forecasts with monthly or quarterly refeaturing.
However, when is reliable dynamic prediction not just? That's about what too. Predictions that lack important signals are too high risk predictions, and predictions that are too large granular You can fill in the important things and waste your team time. Sweetspot focuses on material drivers that really impact your business.
Start streaming.
CFOs are used to looking at revenues for a single figure. However, not all revenue streams are created equal, especially in today's uncertain markets where circumstances can change rapidly. To build reliable forecasts, CFOs must break down revenue and margins into component streams and model accordingly. These streams include:
- Recurring revenue and backlog. This is the most reliable revenue stream, but CFOs need to ensure they are realistic. For recurring revenue businesses, churn and renewal risks should be modeled and analyzed regularly. For project-based companies, CFOs need to track remaining revenue and profits per project (including timing of recognition). When making predictions, they need to separate the “in the bag” (backlog) and the pipeline and what they need to acquire new business.
- Reinducing revenue. As client priorities change and unwarranted, semi-regular revenue streams are less predictable than actual repeat revenue. The CFO should model this stream individually to avoid false stability.
- New projects and M&A revenue. This stream is often overestimated in predictions as well. CFOs may not be able to consider moving to market plans (such as marketing, sales, staffing) due to new projects or acquisitions, and therefore fall victim to excessive optimism about how quickly new revenue streams will increase (risk is common in volatile markets). CFOs should model various outcomes in line with their to-market plans and thoroughly examine the launch of new products and plans of newly acquired companies.
- Ad hoc revenue. Revenues from one-off projects or seasonal increases lack long-term visibility (over 1-2 quarters) and are highly susceptible to fluctuations and overestimation. In comparison, historical trends provide robust (now flawed) baseline, long-term, ad hoc predictionsS should build external market factors, especially if there is no support for pipelines or bookings. cFOS should model this stream conservatively and apply established heuristics such as recent trends and seasonality to avoid overprojection.
Climb the stairs to data heaven.
Sponsors, boards and investors require CFOs to use AI, and revenue forecasts are the best place to start. However, there is a warning. For AI to improve forecasting, two criteria must be met:
- First, the data foundation must be strong. If your business data is messy and incomplete (i.e. unreliable pipelines, data locked in different systems, inefficient or broken processes, incomplete historical data), AI can't add real value. CFOs must start by evaluating and optimizing AI data, including thinking through key internal and external indicators and delay indicators of the business. This allows AI to work with an overall set of both positive signals and past performance data. CFOs must also establish governance to ensure that the quality of their data is maintained on a continuous basis.
- Secondly, AI cannot be everything anywhere at once. Start small. CFOs should focus on products or service lines where gaps (e.g., seasonal air conditioners lines) are forecast, and deploy AI-enabled forecasts at a granular level. However, while AI generates predictions, CFOs need to know how to convert the output into a practical business language. Connect forecasts to specific drivers, KPIs and market factors. Explaining the “why” behind numbers highlights the strategic CFO.
Joanne can predict cross-responsibility, data-driven, scenario-aware, and AI-enabled, allowing modeling to remain reliable even in unpredictable markets.
