The Flawed Math of Long-Term Sales Planning

Why S2S and Sales Velocity Models Fail

For long-term sales planning, most sales organizations reach for one of two trusted frameworks: Stage-to-Stage (S2S) Conversion Rate Models or Sales Velocity Models.

On paper, both look appealing because they are simple to understand and easy to build in a spreadsheet. In practice, however, both are fundamentally misspecified for real-world B2B sales dynamics.

The Two Frameworks

1. Stage-to-Stage (S2S) Conversion Rate Models

You measure historical conversion rates between each sequential stage of your sales process, multiplying those rates together to calculate an overall win rate multiplier (often referred to as the left side of the "bowtie" model). Finally, you multiply that overall win rate by projected deal volume and average deal size:

Win Rate = CR1 × CR2​ × CR3 ​× CR4 ​(for a 5-stage funnel)

Expected sales = Win Rate × Deals × Average Deal Size

S2S models produce a single expected sales figure.

2. Sales Velocity Models

You multiply the number of new opportunities created by your historical average win rate and average deal size, then divide by your average sales cycle length:

Sales Velocity = Deals × Average Win Rate × Average Deal Size​ /  Average Sales Cycle

Sales Velocity produces a single expected sales figure per unit of time (typically per day).

In both models, the output is treated as your expected sales baseline relative to those inputs.

Unique Structural Failures of S2S Models

1. The Conversion Rate Independence Fallacy

S2S models tempt leaders to optimize individual stage conversions in isolation, assuming Stage 1 conversions are disconnected from Stage 3 conversions. In reality, inflating the number of deals passing Stage 1 may be hurting you by simply pushing lower-intent deals downstream—depressing later-stage conversion rates and clogging rep capacity.

2. The Sequential Progression Fallacy

S2S calculations assume deals flow in a clean, linear sequence. Real-world B2B deals frequently start mid-funnel, skip stages entirely, or regress backward. It takes very few non-sequential deals to destroy the model’s accuracy—in an empirical study, just 8% non-sequential transitions caused over a 10x prediction error.

Two Systemic Fallacies Common to Both Models

1. The Deterministic Illusion

Both frameworks produce a single, static point estimate. A single figure—or even an arbitrarily guessed "Low / Base / High" spread—provides zero visibility into probability or variance. They are mathematically incapable of answering core executive questions:

• "What is the probability we fall short of plan by more than 20%?"

• "What are our odds of beating our stretch target?"

2. The Time Blindspot

Standard win rate and conversion calculation methods measure terminal probabilities—whether a deal will eventually convert, assuming you wait forever. They carry zero information about when a deal will close. Velocity models include a "per unit time" element, but that is a steady-state summary. Applying an average sales cycle to an S2S model suffers from the same flaw. This time blindspot over-credits top-of-funnel additions created late in the planning period, while completely ignoring deal age.

The Alternative: Backwards Planning via Properly Specified Models

To build a long-term plan grounded in real B2B sales dynamics, sales leaders must move from misspecified spreadsheet formulas to properly specified empirical models. This requires three structural shifts:

• Time-til-Won Curves: Replace time-blind win rate numbers with empirical win-rate curves that express cumulative win probability as a function of elapsed time after opportunity creation. (Figure 1 below.)

• Probabilistic Distributions: Replace single-point estimates with Monte Carlo simulations (easy to do in a worksheet) to calculate the full distribution of outcomes and expose true downside risk. (Figure 2 below.)

• Backwards Planning: Start with your sales target, quantify your baseline odds of getting there using empirical pipeline behavior. Then stress-test operational levers (deal volume, win rate, deal size) to understand how to shift those odds in your favor.

Figure 1. Time-til-Won curves reflect how cumulative win probability changes as deal age increases. In contrast, traditional win rate formulas apply a static, time-blind ratio that acts as an unreliable estimator of the right edge of the curve. Models built on static rates ignore timing entirely, producing forecasts that are only valid if you wait forever. The same mathematical flaw applies to stage-to-stage (S2S) conversion rates.

Figure 2. Monte Carlo simulations generate probability distributions of potential bookings rather than single point estimates. They estimate the odds of achieving any goal. The horizontal axis represents bookings outcomes; the vertical axis shows relative likelihood. The two curves compare a baseline status-quo forecast (blue) against a strategy-adjusted model (beige). The right (dark blue) vertical line highlights an $8.5M target. Summary statistics below the chart reveal the critical executive context: a status quo 2% chance of beating the target, shifting to 65% once specific operational levers (changes in win rate, deal generation rate, and deal size) are applied.

Stop relying on misspecified S2S and Sales Velocity models that leave you defending execution failures when the starting model was flawed from day one. Build your operational plan around the full distribution of possibilities, know your odds of success, and learn how to shift those odds in your favor.

(If you want to see what a properly specified model looks like in practice, check out our Bookings Planner, which uses Time-til-Won curves and Monte Carlo simulations to power this exact backwards planning approach.)