Demand Forecasting: A Complete Guide

Finance teams are working with more complexity, channels, and tech stacks than ever before. Learn how to modernize your methods.

Benjamin Fox, Product Marketing Senior Analyst, Salesforce

November 20, 2026

Types of demand forecasting

Forecasts sort along three lines: how far ahead they look, how wide they cast, and whether they account for your own plans. Most teams run several at once.

Type Time horizon Primary use case
Short-term Days to weeks Production scheduling, staffing, immediate inventory replenishment
Medium-term Months (1–12) Quarterly budgeting, supplier contracts, resource planning
Long-term Years (1–5+) Capacity planning, facility investment, strategic market decisions
Macro Varies Industry-level demand trends and market size estimation
Micro Varies Product- or SKU-level demand for operational planning
Passive Varies Extrapolates current trends with no assumed interventions
Active Varies Incorporates planned business actions (campaigns, launches, price changes)

Demand forecasting FAQs

Demand forecasting predicts how much customers will buy and when. Demand forecasting estimates how much customers will buy and when. Demand planning is the wider process around it: reviewing the forecast with sales, marketing, and finance, adjusting for promotions and launches, and agreeing on one demand plan. Supply planning then turns that plan into purchasing, production, and inventory decisions.

Quantitative methods use historical data and statistical models, such as time series analysis, moving averages, and regression. Qualitative methods rely on expert judgment, market research, and structured input like the Delphi method or a sales force composite. Most enterprise teams set a quantitative baseline and adjust it with qualitative input for events the data can't capture.

AI models can learn from far more variables than traditional statistical methods, including weather, web traffic, and competitor pricing, and from patterns across thousands of related products. Retrained regularly, they adapt as demand shifts, and many produce a range of likely outcomes, not just one number. Their accuracy depends on clean, connected data.

Teams use spreadsheets for simple forecasts, statistical and machine learning software for advanced modeling, dedicated demand planning and S&OP platforms, and planning features in CRM and ERP systems. Sales Cloud and Revenue Cloud connect that demand to pipeline, quoting, contracts, and billing.

At minimum, you need historical sales or transaction data at the level you want to forecast, such as product, SKU, or region. Stronger forecasts add pricing, promotions, open pipeline, contracts, and external signals like economic indicators and seasonality. Clean and standardize that data before you start modeling.