At ₹2M ARR, your sales pipeline stops being a line chart and becomes a probability puzzle. Pipedrive's default forecast—weighting deal size by win probability—worked fine when your deal size was consistent and your sales cycle was 30–45 days. By ₹2M, you're juggling a ₹50K deal with a competitor to a ₹500K multi-stakeholder deployment. The forecast engine sees both as probability problems. It's not. Weighted probability math assumes independence. Your customer's budget, decision timeline, and stakeholder agreement aren't independent. They cascade. Pipedrive doesn't see that cascade because its standard fields don't capture it. A deal sitting in 'Negotiation' for 12 weeks doesn't look meaningfully different from one there for 2 weeks, yet the forecast treats them as equivalent multiplied by the same win percentage. This post walks you through where Pipedrive's forecast breaks, which custom fields actually predict close rates, and how to rebuild your forecast logic so you can see real pipeline health instead of noise. Why Pipedrive's weighted forecast fails at multi-million-ARR scale Pipedrive's forecast model is brutally simple: Forecast = (Deal Size × Win Probability) summed across the pipeline. That works when your assumptions hold true: Deal sizes cluster within a narrow band (±30% variance). Sales cycle is predictable (±10 days). Win probability is stage-independent (moving a deal from 'Qualified' to 'Proposal' doesn't change the odds materially). Each deal's outcome is independent of others in the pipeline. None of these hold at ₹2M ARR. A ₹50K SMB deal and a ₹750K enterprise deal both living in 'Negotiation' at 60% probability are fundamentally different problems. The SMB deal might close in 3 weeks because the CFO signed off. The enterprise deal might be waiting for legal review—a step that doesn't exist for SMBs. Pipedrive treats them as the same: 0.6 × deal size. Your forecast inflates by the average of both, but the actual risk profile is invisible. Worse: at ₹2M, you probably have deals with overlapping sales cycles. A customer's purchase decision often hinges on whether a competing deal closes first. Your forecast engine doesn't see that dependency. It sums probabilities linearly, which mathematically overstates confidence when outcomes are tied to each other. The 6-month sales cycle problem: when time kills probability weights Pipedrive forecasts best with 4–8 week sales cycles. Everything's fresh. Stakeholders remember the call. Pricing hasn't drifted. By month 6, the deal has usually either closed or stalled hard—but Pipedrive's forecast keeps it at the stage's baseline probability week after week. A deal sitting in 'Proposal' for 2 weeks? 50% win rate is reasonable. A deal sitting in 'Proposal' for 18 weeks? That's not 50%. That's 10%, and the forecast is lying to you. Real pipeline behavior: after 12 weeks in any stage, win probability drops nonlinearly. The customer has gone quiet. You're sending check-in emails. The deal is either fundamentally blocked or it's dead and nobody's said it yet. Pipedrive has no way to detect this because it doesn't track how long a deal has been in each stage relative to your historical cycle time. A deal sitting in the same stage for 6+ months isn't 60% likely to close. It's 85% likely to need a reset conversation or be archived. Custom fields that actually predict close rates Replace Pipedrive's stage-based probability with fields that reveal what your deals actually need: 1. Days in Current Stage (calculated field) Add a custom number field that tracks how many days a deal has lived in its current stage. Pair this with your historical close-time data. If your median deal cycles through 'Proposal' in 14 days and this deal's been there 42 days, that's a leading indicator of stall, not a 50% probability. Action: Any deal exceeding 2× your median stage duration automatically drops to a 'Stalled' flag. Don't wait for a rep to change the stage. 2. Stakeholder Sign-Off Count Track how many decision-makers have approved the deal: budget owner, primary user, legal, procurement, CFO. Deals with 0–1 stakeholder confirmations have vastly different close rates than deals with 3+. Pipedrive's standard fields miss this entirely because they assume approval is binary (in the deal notes, or not). Create a dropdown or number field: 'Stakeholders Approved: [0/1/2/3/4+]'. Your close rate jumps dramatically between 1 and 3. Forecast using that , not the stage. 3. Next Action and Due Date Pipedrive has activity fields, but most reps don't fill them consistently. Create a mandatory dropdown: 'Next Action Required' with options like 'Awaiting customer feedback', 'Internal legal review', 'Waiting for budget cycle', 'Contract review', 'Competitor decision'. Pair it with a 'Next Action Due' date field. Deals where the next action due date is in the past are not 60% likely. They're blocked. Your forecast should flag them separately, not include them at stage probability. 4. Deal Size