How to build a sales forecast
Building a sales forecast is the practical work of turning pipeline into a defensible number. The goal is a repeatable model with explicit assumptions, so the forecast can be inspected, challenged, and improved rather than pulled from a manager's gut.
Short answer
Build a sales forecast by defining the period and metric, pulling clean qualified pipeline, applying stage or historical conversion rates, layering in rep commits and recurring-revenue contribution, then rolling it up into commit, best case, and pipeline categories. Document assumptions and snapshot it so you can measure accuracy against actuals.
Step by step
Define the period, metric, and segments
Decide what you are forecasting (bookings, revenue, ARR), for what period, and at what level of segmentation.
Pull clean qualified pipeline
Gather open deals expected to close in the period, after auditing out stale and mis-staged deals.
Apply conversion rates
Weight pipeline by stage probability or historical conversion so the raw value becomes an expected value.
- Weighted value = deal value x stage or historical probability
- Add recurring-revenue and renewal contribution
- Layer rep commit judgment on top
Roll up into categories
Aggregate into commit, best case, and total pipeline so leadership sees a range with confidence levels.
Document assumptions and snapshot
Record the conversion rates and judgments used, then snapshot the forecast so you can score accuracy and refine the model.
How Ardovo helps
Ardovo builds the forecast from live pipeline with transparent, adjustable conversion assumptions and rolls it into commit, best case, and pipeline automatically. Rook documents the assumptions and tracks accuracy so the model improves each period.
Frequently asked questions
What data do I need to build a sales forecast?
A clean, qualified pipeline with accurate stages and close dates, historical conversion rates by stage, recurring-revenue and renewal data, and rep commit judgment. The cleaner the pipeline and the better your historical rates, the more defensible the forecast.
Should the forecast be a single number or a range?
A range. Roll deals into commit, best case, and total pipeline so leadership sees confidence levels, not one fragile figure. A single number invites false precision and hides the risk in the underlying deals.
How do I make my forecast model defensible?
Document every assumption: the conversion rates, the categorization rules, and the rep judgments. When the model is transparent and snapshotted, you can inspect why it was off and improve it, rather than arguing about a number nobody can trace.