B2B marketing attribution: which channel drives revenue
B2B marketing measures leads and the board asks for revenue. Attribution connects the two when the CRM records where each contact came from and the leaders agree on the model before they see the results, which is what ends the dispute over who generated the deal.
Marketing attribution is the method that links each won deal to the channels and actions that took part in the purchase, so the board knows where to put the next unit of budget. In B2B sales, marketing usually reports on leads and cost per lead while the board asks for revenue; attribution bridges those two numbers when the CRM records where each contact came from and associates the right people with each deal. The fight over who generated the deal ends when marketing, sales and finance sign off on the model before the quarter closes and stop reopening it after they see the results.
In the results meeting, marketing shows that the won account downloaded a guide and attended a webinar, the rep recalls prospecting the director for three weeks and the partner says it made the referral, each one picking the slice that favors its own target. In a survey of 377 users of marketing analytics, Gartner found that analytics influence 53% of marketing decisions, and one third of respondents reported that decision makers cherry-pick data to tell a story that matches the opinion they already held (Gartner, 2022). Among the recommendations in the same release is to set KPIs and metrics before launching a campaign, and not after the data starts coming in, which is the principle I apply to attribution.
What is marketing attribution and which decision does it support?
Each recorded interaction between a person and your company is a touchpoint: the ad clicked, the article read, the form filled in, the event, the SDR’s call, the partner’s referral. The attribution model is the rule that splits a deal’s credit among those touchpoints. With the rule applied to the period’s deals, the company adds up pipeline and revenue by channel and compares them with what it spent on each one.
The decision attribution supports is the budget decision: which channel gets more money and which one needs another quarter before being judged. In RevOps metrics I show how to calculate CAC, LTV and payback for the whole company; attribution breaks those numbers down by channel, and without it CAC by channel is a guess.
Why does marketing measure leads while the board asks for revenue?
The mismatch comes from time. In B2B sales with a four to nine month cycle, a marketing team that waits for revenue to prove results spends two quarters with nothing to show, so it adopts the lead as a short-term indicator. The CFO sees the budget go out every month and wants to know how much came back, and the leads-by-channel report gives no answer, because a channel that produces many contacts can produce few opportunities.
I suggest reading attribution on three levels. The first counts contacts created by channel and shows whether the channel brings people from the right profile, the second counts opportunities created and the third adds up won revenue. The first level shows up within a week, the second within a month and the third after a full sales cycle. Once the board understands that lag, marketing can be held to opportunities in the quarter and to revenue over the year.
What are the attribution models and when should you use each one?
Single-touch models give all the credit to one touchpoint, and multi-touch models split it among several. The percentages in the U-shaped and W-shaped models are the most common market convention, and each company can adjust them to its own funnel.
| Model | How it splits credit | When to use it | Risk |
|---|---|---|---|
| First touch | 100% to the first recorded interaction | Assessing which channels bring new accounts into the base | Ignores what convinced the account after discovery and rewards top-of-funnel channels |
| Last touch | 100% to the interaction before conversion or opportunity creation | Short cycles, with few interactions and a single decision maker | Credits the contact page or the search for the company’s name, and hides what created the demand |
| Linear | Equal parts for every interaction | A first multi-touch step, when the company has no history yet | Gives the third nurture email the same weight as the event that opened the conversation |
| U-shaped | 40% to first touch, 40% to lead creation, 20% split among the rest | Funnels where discovery and lead conversion are the moments the company wants to value | Overlooks what happens between lead and opportunity, which in B2B is usually most of the cycle |
| W-shaped | 30% to first touch, 30% to lead creation, 30% to opportunity creation, 10% split among the rest | B2B sales with a clear opportunity stage and handoff from marketing to sales | Depends on reliable creation dates, and one recording error shifts 30% of the credit |
| Data-driven | An algorithm compares journeys that converted and did not convert and calculates each interaction’s weight | High conversion volume and journeys recorded end to end | Needs more deals than most B2B companies close in a quarter, and the board cannot audit the math |
Which attribution model should you choose?
For most of the B2B companies I work with, the W-shaped model works as the official budget rule, because it credits the three moments marketing and sales recognize as theirs: discovery, lead conversion and opportunity creation. First touch stays as a supporting view, so marketing knows which channels renew the account base. Last touch helps little in committee sales, because it tends to credit the meeting request, and the person who asks for the meeting is rarely the one who discovered the company.
The data-driven model requires volume. A company that closes thirty deals a quarter does not have enough journeys for the algorithm to separate coincidence from effect, and a board that gets a number it cannot recompute tends to distrust it the first time it contradicts its intuition. If the company sells to small and midsize businesses at high volume, the data-driven model can come in after a year of clean records, with the rule-based model running in parallel for comparison.
How to agree on the model before the close
The model goes in writing into the same document that defines MQL, SQL and handoff deadlines, the SLA between marketing, sales and customer success. The marketing and sales leaders and finance sign it, sponsored by whoever decides the budget. The text sets the official model, the window before opportunity creation and how outbound prospecting and referrals show up in the report. The rule holds for at least two quarters, and any change takes effect in the next period, never in the one being closed.
What needs to be right in the CRM before attribution?
Before discussing models, leadership needs to guarantee four records in the CRM, because each model in the previous table depends on at least two of them. Without them, the attribution report looks reliable and carries wrong numbers, and the board only finds out when it cuts the channel that was bringing in revenue.
| Record | What needs to exist | Most common error | Owner |
|---|---|---|---|
| Contact source | A picklist field with the original source, which does not change once filled in, and a second field with the latest source | Source overwritten at each new conversion, or “other” across a large share of the base | Marketing operations |
| Standardized UTMs | A dictionary of allowed values for source, medium and campaign, lowercase and without accents | Each analyst or agency creates its own pattern, and the same channel shows up with five spellings | Marketing, with the agency |
| Contacts associated with the deal | Every participant in the decision associated with the opportunity and the company | Only the signer stays on the deal, and the others’ interactions vanish from the account | Sales rep, with the manager following up |
| Creation dates | Real creation dates for the contact, lead and opportunity, preserved in imports | A spreadsheet import records the upload date, and old contacts show up as new | CRM administration |
Source needs extra care for the channels no tool tracks on its own. In-person events, customer referrals, partners and outbound prospecting need their own values on the list, filled in by whoever creates the contact, or the report credits everything to “direct traffic”. Outbound prospecting goes in as a channel, with the cost of the sales development team in the investment, so marketing does not claim the deal the SDR opened alone. I cover the data entry discipline behind these fields in CRM data quality.
Why does B2B attribution need to look at the account and not only the contact?
In a committee sale, the contact who opens the opportunity is rarely the one who interacted most with marketing. The operations analyst read three articles and downloaded a spreadsheet, the department manager went to an event, the CFO never clicked on anything and she was the one who approved the purchase. If attribution looks only at the contact associated with the deal, the report credits the channel that brought in the CFO, which may be the meeting the rep booked, and erases months of work on the rest of the account.
Account-level attribution adds up the interactions of every contact at the company within a window before opportunity creation. The window comes from the sales cycle: in a hypothetical operation with a median cycle of five months, a six-month window before the opportunity captures the research phase without pulling in interactions from a project the account abandoned two years earlier. For this to work, each contact must be associated with the company, which the CRM does by email domain in most cases, and the decision participants must be on the deal with the role they play, as I describe in the B2B buying committee.
For companies that already work with target accounts, the account view connects with the coverage and engagement indicators in account-based selling. ABM indicators show whether the program is moving forward in each account, and attribution shows which channels took part in the accounts that turned into revenue.
Self-reported attribution: what “how did you hear about us?” reveals
Part of the B2B journey happens where software cannot see, such as an executive’s LinkedIn post or a colleague’s recommendation over lunch. Self-reported attribution asks the buyer how they found the company and records the answer in the CRM, as a complement to the software model. If the software credits organic search and people answer that they heard about the company at an event, search was the path and the event was the cause. The answer has its biases, because buyers remember the last thing they saw, and “Google” often means they searched for the company’s name after hearing about it somewhere else. That is why I read self-reported answers together with the recorded first touch, account by account.
How should you present attribution to the board?
The board needs few columns per channel: how much the company invested, how many opportunities the channel helped create, how much each opportunity cost, how much pipeline and revenue it produced and how long the money takes to come back. Cost per opportunity replaces cost per lead in this conversation, because it brings into the math the leads that never reached sales. In the W-shaped model, attributed opportunities are the sum of the credits the channel received. The figures below are an illustrative example from a hypothetical company, to show how to read the report.
| Channel | Quarterly investment | Attributed opportunities | Cost per opportunity | Attributed pipeline | Attributed won revenue | Median days to win |
|---|---|---|---|---|---|---|
| Events | R$ 120k | 20 | R$ 6,000 | R$ 1.6 million | R$ 300k | 150 days |
| Paid media | R$ 90k | 45 | R$ 2,000 | R$ 900k | R$ 180k | 80 days |
| Content and organic search | R$ 60k | 30 | R$ 2,000 | R$ 1.2 million | R$ 240k | 110 days |
| Partners and referrals | R$ 30k | 10 | R$ 3,000 | R$ 800k | R$ 260k | 70 days |
Looking only at cost per opportunity, the board would cut events and double paid media. The revenue column tells a different story: partners returned more than eight reais per real invested and content returned four, against two for paid media. Events have the largest pipeline and the longest cycle, and cutting them based on a single quarter would judge the channel before its revenue arrives. That is why revenue needs to be read by cohort, adding up what the opportunities created in each quarter closed, even when the close falls in the following year.
Payback time completes the picture. It adds the median days to win to the CAC payback of the customers the channel brings, and tells finance how much cash each channel consumes before giving it back. I take this report to the monthly revenue committee and keep the budget decision for the quarterly review, after at least two periods read with the same model.
What changes for each leader with attribution?
Marketing leadership
Marketing leadership starts defending its budget with opportunities and revenue by channel, and gains an argument to keep long-cycle channels that cost per lead would condemn. In return, it accepts that the channel with the most leads may look bad in the report and that the team’s target moves from volume to opportunities.
Sales leadership
For sales, attribution requires record keeping: associating participants with the deal, filling in the source of the contacts the rep creates and confirming “how did you hear about us” in the first meeting. The payoff is outbound prospecting showing up as a channel, with its own cost and revenue, and the manager seeing which channels bring opportunities that close faster.
Customer success leadership
Expansion revenue needs its own source, separate from acquisition, or the report credits marketing with the upsell that CS ran in an account that was already a customer. With customer events and referral programs recorded as touchpoints, customer success can compete for budget with a measured contribution to renewal and expansion.
A roadmap for leadership to put marketing attribution in place
- Audit the records before the model. Pull from the CRM the percentage of contacts with a source filled in, of deals with more than one associated contact and of campaigns with UTMs that follow the standard.
- Write the dictionary of sources and UTMs. Marketing and sales close the list of channels, including events, partners, referrals and outbound prospecting, and the agency gets the standard in writing.
- Choose the official model and the window. The marketing, sales and finance leaders choose the budget model, the supporting view and the window before the opportunity, and record everything in the SLA.
- Set up account association. CRM administration ensures contacts are associated with the company by domain and makes decision participants mandatory on the deal from the proposal stage on.
- Add self-reported attribution. Put the open field on the form and the picklist field in the first meeting, with a monthly review of the answers.
- Run two quarters without touching the rule. Present the report in the monthly committee, note the questions and adjust the model only when the period turns.
- Take the budget decision to the quarterly review. Compare cost per opportunity, revenue by cohort and payback time by channel, and record the decisions with an owner and a date.
Marketing attribution and L&D: what HR needs to assess
How to recognize the need
The signs show up at the quarterly close: marketing and sales claim the same deal, finance cuts a channel’s budget without knowing how much revenue it brought and the source report shows a large slice of “other”. When the company switches attribution tools and the argument stays the same, the gap lies in record keeping and in reading the report together.
Who takes part
The core group is the marketing, sales and customer success leaders, with finance and the operations or CRM person. Marketing analysts, SDRs and reps join in a second stage, because they are the ones who fill in the source, create the committee contacts and ask the self-reported question. Marketing and sales need to learn together, in an in-company program using the company’s own data, to record source the same way and read the same report; separate groups tend to leave with two versions of the rule.
How to measure
Before the program, use the three percentages from the roadmap audit as the baseline and count how many meetings ended with the model disputed. After a quarter, compare the same numbers, and after two, count how many budget decisions the board made based on the channel report.
If your leaders spend the quarterly close arguing about who generated each deal instead of deciding where to invest, that is the work of the in-company Revenue Operations training. I run the program with marketing, sales, customer success and finance in the same room, on your operation’s source data and deals. The group leaves with the SLA between marketing, sales and CS, with the attribution model already written into it, and with the RevOps metrics matrix that shows pipeline and revenue by channel.
Frequently asked questions
What is marketing attribution?
It is the method that splits each deal’s credit among the touchpoints that took part in the purchase, such as ads, content, events, referrals and prospecting. By adding up that credit by channel, the company compares what it invested with the pipeline and revenue each channel helped produce.
What is the difference between first touch and last touch?
First touch gives all the credit to the interaction that brought the person into the base, and last touch gives all the credit to the interaction before conversion. In B2B committee sales, both leave out most of the journey, which is why I use a multi-touch model as the budget rule and first touch as a supporting view.
What is the best attribution model for B2B?
For most B2B companies, the W-shaped model works well as the official rule, because it credits discovery, lead conversion and opportunity creation. The data-driven model needs a deal volume few B2B companies have. The choice is worth more when the leaders sign off on it before the period closes.
Do I need a dedicated attribution tool?
Market CRMs with built-in marketing automation already offer multi-touch attribution reports. Before buying another tool, check whether source is filled in, whether UTMs follow a standard and whether committee contacts are associated with deals, because no tool fixes those records on its own.
How should outbound prospecting and referrals be handled in attribution?
Treat both as channels, with their own values in the source field and with the cost of the sales development team or the partner program in the investment. That way the SDR gets credit for the account they opened and marketing stops claiming a deal it never touched.
How do you train marketing and sales to use attribution?
The program needs to bring both teams into the same group, with finance and the CRM person, practicing on the company’s real deals. The group learns to record source and committee participants the same way and to read the channel report together. Results show up in how complete the source fields are and in the budget decisions made based on the report.
Want to bring this to your team?
In the in-company Revenue Operations training, leadership designs the SLA between marketing, sales and CS and leaves with the operation's metrics matrix.
More writing on the same subject.

RevOps team: roles, structure and when to hire
Many companies hire a CRM analyst and expect that person to fix process, data, tools and reporting alone. This guide covers the four roles of a RevOps team, how it grows with the company, who it should report to and how HR builds the job description and the training path.
17 min readRead more
Sales tech stack: how to integrate your commercial tools
The company buys one tool per problem and ends up with ten systems that do not talk to each other, retyped data and three versions of the same number. A sales stack works when each layer has a function, an owner and a defined piece of data it returns to the CRM.
18 min readRead more
Sales planning: how to calculate quota and team capacity
A target born from "grow 30%" without checking the funnel and the team turns into pressure by the second quarter. The bottom-up math starts with the customer base, runs through deal size, conversion, cycle, rep capacity and ramp, and shows how many people to hire, when, and how much pipeline the year needs.
18 min readRead more