Pipedrive's forecast math is designed for tight, supervised pipelines: 10–20 deals per rep, disciplined probability marking, monthly closes. Most Pipedrive customers work that way, and the platform works cleanly for them. But the moment you scale to 50+ deals in flight across 3–5 reps—especially in fast-moving verticals like SaaS, recruitment, or marketplace supply—the forecast becomes a noise floor. The problem isn't Pipedrive's code. It's that probability-weighted forecasting assumes consistent rep discipline , and at scale, that discipline fractures. Deal ages slip; probabilities don't update; stale opportunities drift into next month's forecast. By 50 deals, you're forecasting with a signal-to-noise ratio closer to 70:30 than 90:10. This post shows you how to identify that rot and what replaces it. How Pipedrive's forecast math works—and where it breaks Pipedrive calculates forecast as the sum of deal value × win probability across all open deals in a stage. A rep owns a £10K deal at 50% probability; Pipedrive counts £5K toward forecast. This works when: Reps update probability as soon as new information arrives (qualification call, proposal sent, objection raised). Deal age correlates weakly with probability—a 60-day-old deal at 30% probability is genuinely uncertain, not stalled. Stages are narrow, with clear entrance and exit criteria. A deal doesn't sit in 'Negotiation' for three weeks without moving. Deal count is low enough that exceptions surface. With 12 deals, one stale 90-day opportunity stands out. With 50, it drowns in the noise. It breaks when: Reps mark a deal 50% and walk away. The real probability decayed weeks ago, but the number didn't. Deal age is the true signal, not the marked probability. A deal 45+ days in a stage is losing momentum, regardless of what percentage the rep assigned. Stages are sticky. Deals linger in 'Proposal' for 30 days because the prospect is slow to respond, not because the rep is managing it actively. Forecast becomes a sump for every open deal. Reps know the number will be used to set targets or judge their performance, so they keep deals open longer to pad the forecast, driving the signal down further. At 50 deals, condition 4 compounds the others. A rep with 50 open opportunities has weaker incentive to close or kill anything fast. The forecast looks healthy on paper (£250K at 60% = £150K forecast). In reality, 15 of those deals are 60+ days stale, their probabilities haven't moved in 4 weeks, and the true close rate is closer to 35%. The 30% noise floor: what it means for your close A Pipedrive forecast with 30% noise means forecast accuracy swings ±30 percentage points from reality. If your forecast shows £100K, actuals land anywhere from £70K to £130K. That's worse than a coin flip for planning revenue, timing cash, or sizing headcount. More subtly, it hides the actual problem: which deals are real and which are zombies. A rep looks at their Pipedrive board and sees 50 deals with an average probability of 55%. They think they're working a strong pipeline. In reality, 10 of those deals haven't moved in 8 weeks; the rep hasn't followed up; and they'll either close in the next 5 days (at 80% probability) or die (at 5% probability). The forecast has no idea which. This matters because it delays decisions. Finance plans headcount off a £150K forecast. Sales leadership doesn't know that £50K of it is truly dead until day 28 of the month, when reps scramble to close fast deals in the last 72 hours or kill zombies. By then, hiring is already frozen, quota setting is locked, and the month ends with a surprise miss. Audit 1: Find deals older than 45 days in a stage The strongest predictor of forecast rot is deal age, not probability. A deal 50+ days in 'Proposal' is leaking momentum, no matter the rep's written probability. In Pipedrive, pull the full list of open deals and sort by last activity date . Then group by sales stage and look for: Deals 45+ days with no update: These are the forecast rot. A deal that hasn't moved in 6 weeks isn't 50% likely to close; it's in a holding pattern. The rep is waiting for the prospect to call back, or the prospect has gone dark, or the deal was passed to a colleague who hasn't acted yet. Ratio of stale deals to total deals per rep: A rep with 50 deals and 15 stale should have half their forecast stripped. A rep with 20 deals and 2 stale is working a tighter, more current pipeline. Which stages are stickiest: If deals spend an average of 12 days in 'Qualification' but 35 days in 'Proposal', the bottleneck is proposal management, not lead quality. This tells you where to improve the process, not just where the forecast is wrong. Action: Create a Pipedrive report filtered to deals with last activity older than 45 days. Mark every one as 'review' or 'kill' by end of week. This clears the forecast noise and forces the conversation: either revive it with a real action (call, email, proposal refresh) or close it as lost. A rep with