Nobody genuinely expects a marketing forecast to be accurate. What people do expect — reasonably — is that it's honest about what it assumes, and that when it's wrong you can explain which assumption broke. Most marketing forecasts fail that second test, which is why so few leadership teams place any weight on them.
Why most forecasts are useless
They're built by taking last quarter's number and applying a growth rate that matches the target. There's no model underneath, so when reality diverges there's nothing to diagnose — just a gap and an awkward meeting. A forecast that can't be wrong in an informative way isn't a forecast; it's a target with a different label on it.
Build it from the components
Forecast the inputs and let the output fall out. How many opportunities from each source, at what historical conversion rate, at what average value, with what cycle length. Now the forecast is a set of specific claims, each independently checkable. When you miss, you can point at the component that broke — volume was fine but conversion dropped, or the cycle stretched — which is a completely different conversation from 'we came in under'.
- Model volume, conversion and value separately for each source
- Use your own historical rates, not aspirational ones
- State the lag explicitly — this quarter's spend affects next quarter's pipeline
- Show a range, and be clear about what would put you at each end
Publish the assumptions alongside the number
Write down what has to be true: sales stays fully staffed, the conversion rate holds at higher volume, the seasonal dip lands where it did last year. Circulate those with the forecast. It sounds like hedging and it does the opposite — it converts a number people privately distrust into a model people can interrogate. In my experience it also surfaces disagreements early, when they're still cheap to resolve.
Review the misses without defending them
Once a quarter, look at what you forecast, what happened, and which specific component was wrong. Don't defend the number — interrogate the model. After three or four cycles your forecasts get materially better, because you've learned where your assumptions are systematically optimistic. Almost nobody does this review, which is why so many marketing forecasts are no more accurate in year three than in year one.
The goal isn't to be right. It's to be wrong in a way that teaches you something and doesn't cost you credibility.
Want to apply this to your business?
Book a free 30-minute strategy call and we'll diagnose your specific growth blockers together.
Book a Free Call