Pipeline Coverage & Forecasting Playbook

The pipeline you need is set by your win rate: required coverage is about 1 ÷ win rate, so if you win one qualified deal in four, you need about 4× your target in open, qualified pipeline. Then weight that pipeline by your own stage-to-close rates, because coverage tells you whether you have enough, and only a weighted forecast tells you what is likely to close.

Most founders set a revenue target and then hope the pipeline shows up. The math runs the other way. Pipeline coverage is the ratio of your open qualified pipeline to your revenue target for a period: $900K of qualified opportunities against a $300K quarter is 3× coverage. It is a leading indicator. Done right, it flags a gap while there is still time to create or advance deals, not after the quarter has closed.

Every coverage number starts from a win rate, and your own win rate is the real input. There is no borrowed benchmark to reach for: the win-rate figures in circulation disagree with each other because each one divides by a different set of deals, and most do not disclose the sample behind the figure. None of them is a safe substitute for your own close rate.

Win rate is the deals you won divided by the deals that ended, won or lost, over a period long enough to hold a few dozen of them; deals still open stay out. Choose the stage at which a deal counts as qualified, write the rule down and keep it fixed, because that choice alone can halve or double the answer (Figure 1). Then check the rate by value as well as by count. Your target is in dollars, so if large deals lose more often than small ones, the value rate is lower, and it is the one coverage needs.

The relationship is required coverage ≈ 1 ÷ win rate: if you close one in four qualified deals, you need about 4× your target just to break even mathematically. Below one win in four, the curve climbs fast (Figure 2).

The famous “just keep 3× pipeline” advice is that curve at a 33% win rate, and nobody seems to know who first wrote it down. Dave Kellogg, a former CEO of two enterprise software companies, wrote in 2013 that he did not know its history; sales managers were already talking about “the rule of three” before he used his first sales force automation system. Once software made coverage easy to count, he argued, managers pushed every team to 3×, turning it into “a giant self-fulfilling prophecy”^[1]. Treat 3× as a convention, not a finding. If your close rate is materially below 33%, holding 3× coverage isn’t a cushion; it’s a structural miss every quarter.

Two things push the number above 1 ÷ win rate. First, deals slip: a deal dated for this quarter that closes next quarter is a win, but not this quarter’s. So I divide by the win rate times the share of wins that closed in the quarter they were dated for: required coverage ≈ 1 ÷ (win rate × on-time share). That adjustment is my own, not a published formula; it replaces a guessed buffer with a measured one. Second, longer cycles and more stakeholders push the number up, so enterprise coverage naturally runs above mid-market coverage in the same business, which is why Step 3 splits the pipeline by segment.

If you have too few closed deals to measure, plan on a win rate of 15% to 25% for an early-stage B2B team, which means 4× to about 7× coverage. That is my planning assumption, not a benchmark; replace it with your own rate as soon as you have one.

A coverage ratio is only as good as the pipeline it counts, and the cause of a bad forecast is rarely the model. It’s the data feeding it. Data quality is a common weak point: in Gartner’s 2020 State of Sales Operations survey, only 47% of respondents believed their organization had high-quality data, and only 45% of sales leaders and sellers had high confidence in their organization’s forecasting accuracy^[2]. Before you divide by the target, take out three kinds of deal:

1. Unqualified deals. Coverage built on unqualified deals is fiction. A deal counts only once it meets the rule you used for the win rate in Step 1 (the ICP and Account-Scoring Playbook sets the fit half of that rule); adding more bad pipeline inflates the ratio without improving the outcome.
2. Stale deals. Anything untouched for about 60 days, or older than about twice your median sales cycle, shouldn’t count toward this period. Both cut-offs are my rules of thumb, not published thresholds.
3. Deals that cannot close in time. A deal counts toward the period only if its close date falls inside it and its stage can still close in the days left; Step 7 shows the check.

Then split what is left by segment. One blended number hides the truth: inbound and outbound, mid-market and enterprise don’t convert the same way, so check each segment’s coverage against its own win rate, unless you have too few deals to split, in which case see Step 4.

Coverage tells you whether you have enough pipeline. A forecast tells you what’s likely to close. They are different questions, and conflating them is how teams get surprised: two companies at the same coverage can land very different amounts, because one qualifies harder or has more of its deals at late stages. Coverage is the volume check. The weighted forecast is the reality check. You need both.

A weighted forecast multiplies each deal’s value by its stage probability and sums the results: the realistic expectation, not the best case. A $40K deal at a stage that closes half the time contributes $20K; the same deal at a stage that closes a quarter of the time contributes $10K. That difference is the whole point.

The probabilities have to come from your own history: for each stage, the share of deals that reached it and were eventually won. A forecast built on optimistic, made-up probabilities is just a wish with a spreadsheet around it, and a CRM’s defaults are made up for you. HubSpot’s default sales pipeline, for example, sets 20% at “Appointment scheduled”, rising to 90% at “Contract sent”, and multiplies each stage’s total by its probability to show a weighted amount^[3]. Test any signal you want to weight the same way, such as a second contact engaged in the deal: compare the win rate of deals with and without it before you give it extra weight.

Stage rates say whether a deal will be won, not when. If some of your wins slip, multiply the weighted total by your on-time share from Step 2. The free Pipeline Forecast and Coverage Calculator computes the weighted forecast, the coverage ratio, the gap to target and how much new pipeline you’d need, using the same math.

What if you don’t have stage probabilities yet?

Step 4 assumes you can count how many deals reached each stage and how many of those were won. Most seed teams can’t yet: either the CRM stages changed twice, or there are a dozen closed deals in total. You can still forecast. You just forecast differently, and you say how rough it is.

Use one blended win rate for coverage, and a deal-by-deal call for commit. Take your win rate from every deal that ended in the last four to six quarters: won divided by won plus lost, counted from the qualified stage, the same way as in Step 1. Step 3 warned that a blended number hides the difference between segments; here it is the fallback, because you don’t have enough deals to split.

Apply the blended rate to the open deals you can’t vouch for, then judge each remaining deal by name. A deal goes into commit only on deal-level evidence: a verbal yes, an agreed price and a named signer. The CRM stage doesn’t decide it. That is the same test as the Step 5 Do this box, applied without a probability table.

Treat the win rate as a range, not a point. Three wins out of 12 ended deals is 25%, but the 95% range (Wilson interval) runs from about 9% to 53%. At the low end you would need about 11× coverage; at the high end, about 2×. Plan on the low half of that range until the sample grows; Step 2’s 15% to 25% planning assumption is a sensible place to sit.

With seven deals, the weighted number is a guide, not a promise. In practice you will land on one of a few outcomes, as the common question on large deals explains. Read the forecast deal by deal, and use the coverage line for one decision: whether to build pipeline now. Here it says build: at 1.2× against about 4×, pushing the deals you have will not reliably close a $78,750 gap.

Then add two bounding numbers. Commit is your near-certain deals: the floor. Best case is the ceiling. Write down what each means, because tools use the words differently. The calculator counts as commit every deal at a stage with a probability of 75% or more, at full value, and counts every open deal closing as the best case; the tracker computes commit the same way. In Salesforce, forecast categories run from Pipeline through Best Case and Commit to Closed, and a cumulative best-case forecast counts Best Case deals together with the Commit and Closed ones (and Most Likely, if used), leaving early-stage Pipeline deals out^[4]. A CRM’s best case can therefore sit well below the ceiling.

Report the three numbers together, against the target. In my view, plan on the expected number, and commit spending you cannot reverse, such as a new hire, only against commit.

When the forecast falls short of the target, you have four levers: create more qualified pipeline, lift your win rate, increase deal size, or accelerate velocity, so that more deals close inside the period. The funnel that feeds the first lever, from lead to qualified opportunity, is the subject of the Lead Management Playbook.

The right lever depends on timing. Early in the period, the answer is usually to build pipeline, but only qualified pipeline, because stuffing the funnel with marginal deals makes forecasting worse, not better. How early is set by your median sales cycle: once fewer days are left in the period than the cycle takes, a newly qualified deal will, at its median pace, close next period. Later in the period, early-stage pipeline can’t close in time, so the lever shifts to pushing late-stage deals and lifting conversion on what you already have. I work through that trade-off on a full pipeline in Our Pipeline Is Full and Nothing Is Closing.

Pipeline coverage only works as an early-warning system if you actually look at it on a cadence. In my view, weekly is the right one: a quarter has 13 weeks, and a monthly review gives you three chances to act, the first after the build window has usually closed. A simple operating rhythm:

• Weeks 1 to 3: measure full-pipeline coverage against the whole target.
• From the end of the build window: shift to to-go coverage, the pipeline that can still close against the target still to win. A deal can still close if its stage’s median days to close fit in the days left (Figure 7).
• Always: segment by motion, deal size and stage. A healthy blended ratio can hide a starving enterprise segment propped up by low-value inbound.

At the end of each quarter, compare the forecast you made on its first day with what closed. Treat your own last four quarters of forecast-to-actual as the only accuracy benchmark that applies to you: it is the one measurement built on your deals, your stages and your slippage. That’s the difference between “we need more leads” and “we need 12 more qualified mid-market opportunities this month or the quarter is at risk.” One is a feeling. The other is a decision you can defend.

What changes by stage?

What mistakes should I avoid?

• Holding 3× because everyone does. Use 1 ÷ your win rate, adjusted for slippage (Step 2).
• Counting every open deal. Take out unqualified and stale deals before you divide (Step 3).
• Forecasting on the CRM’s default probabilities. Use your own stage-to-close rates (Step 4).
• Reporting one number. Show commit, expected and best case together (Step 5).
• Adding new pipeline late in the quarter to save it. After the build window, advance the deals you have (Step 6).
• Reading only the blended ratio. Check each segment against its own win rate (Steps 3 and 7).