Agent forecasting is harder but not impossible
Forecasting revenue from commission only agents is more challenging than forecasting direct sales. You have less visibility into agent activity, less control over their pipeline, and less certainty about their continued engagement. But with the right data and methods, you can build forecasts that are reliable enough for business planning.
The data you need
Historical performance
Analyse at least three to six months of agent channel performance. What is the average revenue per active agent per month? How many agents are consistently active? What is the typical deal cycle length?
Pipeline data
What is the current total pipeline value across all agents? What is the historical conversion rate from pipeline to closed deals? These two numbers give you a basic forecast: pipeline value multiplied by conversion rate equals expected revenue.
Agent capacity
How many agents are currently active? How many new agents are expected to join in the forecast period? What is the typical ramp time for new agents? Active agents at full capacity generate different revenue than new agents in their ramp period.
Forecasting methods
Bottom up by agent
Estimate revenue for each agent based on their individual pipeline, historical performance, and trajectory. Sum these individual forecasts for a total channel forecast. This is the most accurate method but the most time consuming.
Average productivity model
Multiply the number of active agents by the average revenue per agent. Adjust for new agents (who produce less during ramp) and for seasonal patterns. This is simpler but less precise for individual agents.
Pipeline weighted model
Apply probability weights to each deal in the pipeline based on its stage. Early stage deals might have a 10% to 20% probability. Late stage deals might have a 70% to 80% probability. Sum the weighted values for a probability adjusted forecast.
Adjusting for agent specific factors
Agent tenure
Experienced agents with established networks produce more predictable revenue than new agents. Weight your forecast toward experienced agents and apply higher uncertainty ranges to newer ones.
Seasonal patterns
If your sales have seasonal patterns, apply seasonal adjustments to your forecast. A forecast that assumes flat demand across all months will be wrong during peak and trough periods.
Agent churn
Factor in expected agent turnover. If you typically lose 20% of agents per year, your forecast should account for this attrition and the time required to replace departing agents.
Forecast accuracy tracking
Compare forecast to actual
Every month, compare your forecast to actual results. Track the variance. Over time, you will learn where your forecasting model is consistently too optimistic or too pessimistic, and you can calibrate accordingly.
Identify systematic biases
If your forecasts are always 20% high, you have a systematic optimism bias. If they are consistently low, you may be underweighting pipeline quality or agent capability. Adjust your model to correct for observed biases.
Communication
Share revenue forecasts with your leadership team, finance team, and anyone involved in resource planning. Be transparent about the confidence range. "We forecast $150,000 from the agent channel next quarter, with a range of $120,000 to $180,000" is more useful than a single number.
On Zepys, pipeline and performance data is tracked in real time, providing the inputs you need for accurate forecasting without requiring manual data collection from agents.