Every sales team has an estimate. A projected MRR. A number on a deal that says this is what this customer will be worth.
We wanted to know: how often is that number right?
We took a real 3PL company’s data — 473 CRM deals and 77 months of actual revenue by customer — and matched them together. 246 deals connected to real revenue records. After excluding batch-imported legacy data with unreliable close dates, the clean analysis covered 173 customers.
Here’s what we found.
1. Close to First Bill: Median 44 Days
How long from the day a deal closes to the first month with revenue?
| Bucket | % of Customers | Cumulative |
|---|---|---|
| 0–15 days | 19% | 19% |
| 16–30 days | 15% | 34% |
| 31–60 days | 23% | 57% |
| 61–90 days | 16% | 72% |
| 91–120 days | 8% | 80% |
| 121–180 days | 7% | 88% |
| 181+ days | 12% | 100% |
57% of customers generate first revenue within 60 days. But 20% take longer than four months. That tail matters — it’s cash that’s not arriving when the forecast says it will.
The CRM’s “Expected Onboard Date” field was essentially unusable: 97% of records showed a default null value. Only the most recent 14 deals had real projected dates. That’s a process gap worth fixing — if you’re going to forecast, you need to know when the clock starts.
2. Ramp to MRR: Fast When It Happens
For customers who reached 80% of their estimated MRR, the median time was 2 months. The ramp itself isn’t the problem.
The problem is that 28% of customers never reach 80% of their estimate at any point in their entire tenure. They onboard, they bill, but the revenue never gets to the level the salesperson projected.
3. MRR Estimate Accuracy: The Big Finding
Median accuracy: 65% — the typical deal delivers about two-thirds of what was estimated.
| Accuracy Band | % of Customers |
|---|---|
| Under-delivered (<80% of estimate) | 54% |
| On target (80–120%) | 13% |
| Over-delivered (>120%) | 32% |
Over half of customers generate less than 80% of the MRR the salesperson estimated. Only 13% land in the “on target” band. The 32% that over-deliver are real — some customers significantly outperform — but the median tells the story.
If this company is using sales estimates to project revenue, they’re systematically over-forecasting. A haircut of 30–35% on pipeline estimates would bring projections closer to reality.
This isn’t a sales problem per se. The estimates might be reasonable at the time of signing. But customer behavior after onboarding — shipping fewer units than expected, seasonal patterns, slower ramp — erodes the number. The question is whether the forecasting model accounts for that erosion or pretends it doesn’t exist.
4. How Long Do Customers Stay?
37% of the clean set is still active. 63% have churned.
For churned customers:
| Tenure | % of Churned |
|---|---|
| 1–3 months | 6% |
| 4–6 months | 12% |
| 7–12 months | 28% |
| 13–18 months | 17% |
| 19–24 months | 18% |
| 25–36 months | 12% |
| 37+ months | 7% |
Median tenure for churned customers: 13 months.
The dangerous zone is months 4–12 — that’s where 40% of churn happens. Customers who survive year one have better odds of sticking around, but the second year still sees significant attrition. Very few become truly long-term: only 19% of churned customers lasted more than two years.
What This Means
This analysis points to three operational gaps:
Forecasting needs a reality adjustment. Sales estimates are running 35% hot. Either recalibrate the estimation process or apply a systematic discount to pipeline numbers. Both would improve forecast accuracy.
The first year is the retention cliff. If you’re going to invest in customer success, months 4–12 are where the intervention has the highest leverage. By the time a customer churns at month 14, the decision was probably made months earlier.
Onboarding speed varies wildly. The 20% of customers who take 4+ months to generate first revenue represent delayed cash flow and — likely — a higher churn risk. Tightening the onboarding window would compress the revenue curve and might improve retention.
Methodology
- Revenue data: Monthly Sales by Customer Summary (QBO), Jan 2020 – May 2026
- Sales data: CRM export with customer name, estimated MRR, and close date
- Matching: Normalized name matching between CRM and revenue records (246 of 473 matched)
- Steady-State MRR: Average monthly revenue from month 4 onward (skips ramp period)
- “Still Active” defined as having revenue in any of the last three months of data
- Batch-imported legacy deals (with identical close dates suggesting CRM migration, not actual close) were excluded from summary statistics but included in the detail workbook