How to forecast more accurately, without leaning on shaky probabilities
Forecasting is the heartbeat of growth planning. But if you’re still using probability-based models (5% here, 50% there) without enough data to back them up, your forecast is more fiction than fact.
In this on-demand session, Marc DiGiorgio breaks down why so many forecasting models miss the mark, and how tech companies can shift to smarter, more qualitative forecasting for better predictability and boardroom confidence.
Watch the webinar or read on for some of the key takeaways.
Why probability-based forecasting breaks down
If you’ve ever forecasted based on deal stages, such as assigning 5%, 25%, or 50% probabilities, you know the logic: you average it all out and hope the numbers net out. But in most tech companies, the dataset just isn’t big enough to train meaningful probabilities. And worse, if sales teams aren’t diligent about updates, you’re building a forecast on stale data and guesswork.
Marc puts it plainly: “Probability-based forecasting is guaranteed to be inaccurate.”
Qualitative forecasting gives you something better to stand on
Instead of fuzzy probabilities, qualitative forecasting uses forecast categories, like Committed, Best Case, or Pipeline, to build a range-based forecast that’s actually rooted in real sales activity. When combined with deal value (like ARR or total services), this approach gives finance teams a more realistic picture of what’s likely to land.
Marc suggests creating two core scenarios:
-
Base forecast: Includes closed/won and committed deals
-
Stretch forecast: Adds in best case deals that are showing strong signs of closing
This way, your CFO can see both conservative and optimistic views—backed by actual behavior in the pipeline.
What makes a “Best Case” deal?
Just because a deal is technically in the pipeline doesn’t mean it belongs in your stretch forecast. A best case opportunity should show real signals:
-
Next steps are clearly defined
-
There’s recent activity and engagement
-
You’ve identified stakeholders and competitors
-
It’s not stuck in the early stages
-
It hasn’t been pushed quarter after quarter
-
There’s a quote or proposal in play
Deals that linger too long in early stages, or show no recent movement, don’t belong in your best case.
Clean up your funnel to clean up your forecast
Old, stagnant, or never-was deals? They’re skewing your win rate and adding noise to your pipeline. Marc advises sales leaders to regularly sweep out aging deals, especially those way past average sales cycle length, and reset the opportunity structure.
That cleanup doesn’t just improve forecast accuracy. It also sets the stage for more meaningful coaching.
Can AI help? Eventually (but don’t skip the human step)
Yes, platforms like Salesforce and Microsoft have had AI scoring tools for years. But in Marc’s view, they’ve often fallen short due to poor or incomplete data. AI that only looks at the opportunity table and not the activity or relationship data doesn’t tell the full story.
His advice: Don’t wait for AI to fix your forecast. Use regular one-on-one funnel reviews with reps to challenge assumptions, identify real signals, and coach your team toward better execution.
“It’s a great opportunity to help them grow—and make the next opportunity that much better.”
Watch the webinar to see how Marc builds base and stretch forecasts using TekStack, and what indicators he looks for to separate wishful thinking from real pipeline.
Recent posts
