ARR forecasting is often treated as a straightforward prediction problem, but the hardest part is rarely producing a number. Different customer segments, renewal cycles, expansion patterns, churn behavior, and pricing changes can all create very different revenue dynamics, and combining them into a single model can quickly hide the signals that matter most. In this session, Sumathy will explore how to build a forecasting system that preserves that business structure by forecasting distinct segments independently, reconciling them into a consistent topline view, and creating a model stakeholders can actually understand and trust. She’ll also look at why interpretability and uncertainty matter just as much as raw forecast accuracy. From using approaches such as ARIMA and AutoARIMA to scale forecasting across different parts of the business, to presenting ranges rather than a single point estimate, the session will show how forecasting can become a shared decision-making tool for Finance, RevOps, Customer Success, Product, and leadership - helping teams understand what is changing, where risk is emerging, and what is actually driving the numbers.