Why marketing and sales pipeline numbers don't match
Same CRM, same definitions of a lead and an opportunity, and still two answers to "how much pipeline did marketing bring?". Usually nobody made a mistake. Sales reads the pipeline deal by deal, each deal counted whole. Marketing reads it through an attribution model, which splits each deal across the touches that came before it. The two readings answer different questions, and only one of them is built to give totals.
How an attribution model splits a deal
A touch is any recorded activity by a person linked to the opportunity: a webinar attended, an email clicked, a trial started, a demo requested. Depending on the company, the people counted are the contacts on the opportunity or everyone at the account. Three touches can come from one person or from three.
A U-shaped model (also called position-based), common in B2B, gives 40% of the deal to the first touch and 40% to the last touch before the opportunity is created. The remaining 20% is split across the touches in between. A deal with a single touch gives it 100%.
Take two deals from the same quarter (illustrative figures):
- Trial sign-up: 100% of deal A
- Paid search ad, then a guide: 40% of B
- Blog article found through search: 10% of B, $5,000
- Newsletter: 10% of B, $5,000
- Webinar, the last touch: 40% of B
The ad brought a new person in and gets $20,000, as does the webinar, the last touch before the opportunity. The newsletter reached someone already in the database and gets $5,000. The total matches the CRM. Apart from the trial, which was deal A's only touch, no channel's figure matches a deal.
Why a sales leader doesn't find those numbers in the CRM
The credit is split, the deals are not. A sales leader looking for the deals behind "paid search: $20,000" finds one deal worth $50,000, which also involved an article, a newsletter and a webinar. The figure is correct, but it matches nothing they can open in the CRM. Over a quarter with dozens of deals, each channel's figure adds up fractions of many deals.
Touches with no link to the deal. An account executive has been working an account for weeks, with calls and a workshop for the team. Meanwhile, someone else at the same company, not involved in the project, attends a webinar. When the opportunity is created, a model that counts touches from everyone at the account gives that webinar 40% of a deal the sales team built. The data is right: the webinar happened, at that account, before the opportunity. The link to the deal is missing.
Two teams, one deal. Marketing, sales and partners each report their contribution with their own rule. Put next to each other, their shares can add up to more than the whole pipeline, and each team can point to deals that prove its case.
Pipeline reporting both teams can trust
- Take totals from the CRM. Pipeline, deals and revenue come from the CRM. Attribution compares channels: read it row by row, or as orders of magnitude.
- Name the rule, not just the label. "Sourced", "touched" and "influenced" mean different things from one company to another. Write the rule under each figure: which touches count, over what window, and how the credit is split. That is what a single source of truth should hold: shared definitions, not a single number.
- Decide whose touches count. Counting everyone at the account catches buying committees, but also activity with no link to the deal. Counting only the contacts on the opportunity is stricter, and depends on sales adding them.
- Record sales activity as touches. Calls and meetings logged in the CRM can count as touches too, so the model sees the work done on the account.
- Show the deals behind a channel's figure. A list of the deals that make up a channel's credit, with each deal's share, answers "where is that $20,000?" faster than any explanation.
Done by hand, this means rebuilding every quarter the list of deals and touches behind each channel. Funnelsight shows the CRM total next to the figures from the attribution template your team picks, with the rule written under each one. When the two differ, it says why: deals with no recorded touch, touches outside the window, or dates that differ (attribution follows touch dates, sales reports follow creation or close dates). Meetings and calls recorded in the CRM count as touches like any other, and next to a channel's share, Funnelsight shows how many deals it touched. The CRM stays the reference for pipeline and revenue, and Funnelsight doesn't calculate compensation. The same view shows the recurring revenue each source brings and how fast it pays back what it cost.
See your CRM totals and your attribution side by side.
Start free trialWhat do "sourced", "touched" and "influenced" mean?
It depends on the company. "Sourced" can mean the first recorded touch, the lead source in the CRM, or the team that opened the opportunity. "Touched" and "influenced" vary just as much, mostly in which touches count and over what window. Before comparing figures, check the rule behind each one.
Does attribution prove a channel caused a deal?
No. An attribution model shows which touches came before a deal. It can't tell whether the deal would have happened without one of them. It's a way to compare channels, not proof of cause. Tests with holdout groups get closer to cause, but few B2B teams have the volume to run them often.
Does this replace our CRM?
No. Funnelsight reads from your CRM and the other tools you already use. The CRM stays the reference for pipeline and revenue.