Accurate sales forecasting is essential for business planning, resource allocation, and investor communication. SalesSheet.ai combines traditional weighted pipeline methodology with AI-powered predictive analytics to generate forecasts you can actually trust. This guide explains how forecasting works and how to leverage both automated insights and manual adjustments for maximum accuracy.
Prerequisites
- Active deals in your pipeline with values and close dates
- Configured stage probabilities (see Understanding Deal Stages)
- At least 30 days of historical deal data for AI predictions
Understanding Weighted Pipeline Forecasting
Weighted pipeline forecasting multiplies each deal's value by its stage probability to calculate expected revenue. This provides a more realistic forecast than simply adding up all deal values.
For example, if you have three deals:
- Deal A: $100,000 in Qualified stage (25% probability) = $25,000 weighted value
- Deal B: $50,000 in Proposal stage (40% probability) = $20,000 weighted value
- Deal C: $75,000 in Negotiation stage (60% probability) = $45,000 weighted value
Your total weighted pipeline value is $90,000, which represents your expected revenue from these three deals. This is significantly different from the $225,000 total if all deals closed, giving you a more realistic planning number.
SalesSheet.ai calculates weighted pipeline automatically based on your stage probabilities. Access your forecast by navigating to Reports > Forecast or clicking the Forecast widget on your dashboard.
You can view forecasts by time period (this month, this quarter, this year), by pipeline, by owner, or by any custom field. This flexibility helps different stakeholders see the forecast views most relevant to them.
Using AI-Enhanced Win Probability
While stage-based probabilities are useful, SalesSheet.ai's AI analyzes individual deal characteristics to generate deal-specific win probabilities that are often more accurate than stage averages.
The AI considers dozens of factors when calculating win probability for each deal:
- Historical win rates for similar deal sizes in your CRM
- Company size, industry, and geographic factors
- Engagement level: email response rates, meeting attendance, document views
- Deal velocity: how quickly the deal has progressed compared to historical norms
- Stakeholder involvement: number and seniority of engaged contacts
- Time since last activity: deals with recent engagement score higher
- Competitive dynamics: deals with known competitors score differently
- Seasonality patterns: certain products or industries have seasonal buying cycles
Each deal displays both the stage probability and AI-calculated probability in the deal detail panel. When they differ significantly, click "Why?" to see which factors are influencing the AI's assessment.
In forecast reports, you can toggle between "Stage-Based Forecast" (using stage probabilities) and "AI-Enhanced Forecast" (using deal-specific probabilities). Most teams find the AI forecast becomes more accurate than stage-based after 60-90 days of data collection.
Configuring Forecast Time Periods and Views
Navigate to Reports > Forecast to access the main forecasting dashboard. Here you can customize which deals contribute to your forecast and how results are displayed.
Time period filters include:
- This Month: All deals with expected close dates in the current calendar month
- This Quarter: Deals expected to close in the current quarter (Q1: Jan-Mar, Q2: Apr-Jun, Q3: Jul-Sep, Q4: Oct-Dec)
- This Year: Deals with close dates in the current fiscal year
- Next 30/60/90 Days: Rolling windows from today's date
- Custom Date Range: Specify exact start and end dates for your forecast period
Your forecast can be segmented by:
- Owner: See individual rep forecasts and roll up to team or company totals
- Pipeline: Separate forecasts for different pipelines (new business vs. renewals, for example)
- Product Category: Forecast by product line or service type
- Region: Geographic forecasts for territory planning
- Deal Source: Track forecasted revenue from different lead sources (inbound, outbound, partner, etc.)
Save custom forecast views for quick access. For example, create saved views for "Q1 New Business by Rep," "Annual Enterprise Forecast," or "Next 90 Days - Product A."
Reading Forecast Reports and Insights
The forecast dashboard shows several key metrics and visualizations to help you understand your expected revenue:
- Total Pipeline Value: Sum of all deal values without probability weighting (best case scenario)
- Weighted Forecast: Probability-adjusted expected revenue (most likely scenario)
- Commit Forecast: Conservative estimate based on deals at 60%+ probability (high confidence)
- Closed Won to Date: Revenue already achieved in the forecast period
- Gap to Goal: Difference between your quota/target and current weighted forecast
- Forecast Trend: How your forecast has changed over the past 30 days
The waterfall chart breaks down your forecast by stage, showing how much expected revenue sits in each part of your pipeline. This helps identify if you're too dependent on early-stage deals or if your pipeline is healthy.
Deal health indicators flag at-risk opportunities within your forecast. Look for red or yellow indicators on high-value deals and investigate why they're flagged. Common risk factors include stalled deals, approaching close dates without recent activity, or low engagement scores.
The AI Insights panel highlights trends and anomalies: "Your Q1 forecast has increased 15% in the past week due to 3 large deals entering Negotiation stage" or "Warning: 40% of your forecast depends on deals closing in the next 10 days."
Tracking Forecast Accuracy Over Time
Understanding how accurate your forecasts have been historically is crucial for improving future predictions and building confidence in your numbers.
Go to Reports > Forecast Accuracy to see how your predicted revenue compared to actual closed revenue for past periods. The accuracy report shows:
- Forecast vs. Actual: Side-by-side comparison of what you predicted and what actually closed
- Variance Percentage: How much your forecast was off (over or under) as a percentage
- Trend Over Time: Whether your accuracy is improving or declining over months/quarters
- Accuracy by Forecaster: Which team members consistently forecast accurately vs. optimistically/pessimistically
- Stage-Specific Accuracy: Whether your stage probabilities match reality or need adjustment
Forecast accuracy above 90% is considered excellent, 80-90% is good, and below 80% indicates a need for process improvement or probability recalibration.
If your forecasts are consistently too high, your stage probabilities are likely too optimistic, or reps may be advancing deals prematurely. If forecasts are consistently too low, you might be too conservative with probabilities or not accounting for deals that close faster than expected.
Review forecast accuracy monthly and adjust stage probabilities, required fields, or sales process accordingly. SalesSheet.ai can auto-adjust stage probabilities based on historical performance if you enable this feature in Pipeline Settings.
Pro Tip: Forecast Categories for Commit, Best Case, and Pipeline
Train your team to categorize deals into three forecast buckets: Commit (90%+ likely to close this period), Best Case (50-90% likely), and Pipeline (everything else). This gives leadership multiple scenarios for planning. Ask the AI "Show me my commit forecast for this quarter" to see only your highest-confidence deals.
What to Expect
After implementing robust forecasting practices, you'll see:
- Real-time forecast updates as deals progress, get created, or close
- Automated weekly forecast emails to stakeholders showing changes from prior week
- AI alerts when forecasts are trending significantly up or down
- Improved accuracy over time as the AI learns from your team's historical patterns
- Better business planning with reliable revenue predictions 30-90 days out
- Identification of coverage gaps when forecasts fall short of goals
Most teams achieve 85%+ forecast accuracy within 90 days of consistent usage and probability calibration.