How to Build a Pipeline Model Your CFO Believes

A pipeline model is not a forecasting exercise for its own sake. It is the document that turns a marketing budget from an expense into a calculation, and it is what stops the conversation where finance asks what happens if they cut the number by 30 percent and nobody in the room can answer.

This is the model we build with every B2B SaaS client in the first month of an engagement, before any channel decisions in the demand generation playbook get made. It takes half a day to construct and it changes how marketing is discussed for the rest of the year.

Why 2026 models built on 2025 assumptions are already wrong

Before any arithmetic, one calibration point. Conversion assumptions across B2B have moved sharply, and a model that reuses last year's win rate will understate required pipeline by a third or more.

Read what that does to the chain. At a fixed revenue target, moving win rate from 29 to 19 percent raises required opportunities by 53 percent. If the budget stayed flat, the plan was already missing before the year began. This single number is the most common reason a well built model produces a wrong answer.

The core chain

Six numbers, in this order. Every one of them is an assumption, and writing them down is what makes them correctable.

Every stage formula, written out

Work each one from the bottom of the funnel upwards. Building forwards from traffic is how teams end up presenting a plan that produces exactly the revenue they hoped for.

Three worked models at different ACV bands

The same chain produces very different operating advice depending on contract value. These are complete, self contained models you can copy and swap your own numbers into.

A blended model averages a 12,000 dollar motion with a 180,000 dollar motion and produces a number that describes neither.

How much pipeline coverage do you need?

Coverage is the multiple of pipeline value you need relative to the target, and it exists because deals slip, stall, and die. Three times is the common default. The right number for you is derived from your own history, not from a benchmark.

• Short cycles under 60 days with stable win rates: two and a half to three times.
• Typical mid market motion: three times.
• Enterprise, multi stakeholder, long cycles: three and a half to four times.
• New segment or new product with no history: four times, and revisit after two quarters.

There is also a timing dimension people forget. Pipeline created in the last month of a quarter with a 90 day cycle contributes to next quarter, not this one. Model creation date and expected close date separately or you will consistently overstate the current quarter.

Segment the model or it will lie to you

A blended model averages a 15,000 dollar self serve motion with a 200,000 dollar enterprise motion and produces a number that describes neither. Build the chain separately for each segment that has a genuinely different contract value, win rate, or cycle length, then sum them.

The moment the model is segmented, allocation decisions become obvious. If enterprise cost per opportunity is 12,000 against a 180,000 contract value, that is a better ratio than mid market, and the budget should reflect it even though the absolute number looks alarming.

Capacity planning: the constraint nobody models

Marketing builds a model that produces 900 opportunities. Sales has seven reps. Nobody checks the arithmetic in between, and two quarters later the complaint is lead quality when the actual problem was that each rep received more opportunities than a human can work.

CAC payback: the number that keeps you funded

Use fully loaded cost. Media spend alone flatters the number badly; include salaries, tools, agency fees, and content production. The version finance believes is the version that includes everything.

Stress test before you present

Present three scenarios rather than one. It changes the meeting from a debate about your optimism to a discussion about risk.

1. 1

   Base case

   Current conversion rates held flat. This is the plan.

2. 2

   Downside

   Win rate down 20 percent, cycle length up 30 percent, cost per opportunity up 15 percent. Show what still lands.

3. 3

   Upside

   One specific improvement, such as speed to lead or a conversion path fix, and what it unlocks. Attach the investment it requires.

Assumptions that most often prove wrong

1. Win rate borrowed from a benchmark instead of your own closed won data.
2. Win rate carried over from last year, when the market average fell ten points.
3. Cycle length measured from opportunity creation when sales measures it from first contact.
4. Cost per opportunity assumed constant while scaling spend, when it rises as you exhaust the best audiences.
5. Ignoring capacity: pipeline nobody has time to work is not pipeline.
6. Treating average contract value as stable while the sales team drifts down market to hit deal count.
7. Counting the same opportunity in two channels because attribution double credits it.
8. No seasonality, in a business where two months of the year are structurally quiet.
9. Rep ramp ignored, so a hire in month seven is modelled as a full year of capacity.

Keep it alive

A model that is rebuilt once a year is a slide. Update the conversion rates monthly from actual data, review the assumptions quarterly with sales in the room, and record what you changed and why. Over three or four quarters the model becomes the most accurate forecasting instrument the company has, and the marketing budget stops being the first line anyone reaches for.