Marketing Attribution to Revenue: Which Channel Produced the Customer?
Traffic and signups are easy to see; revenue by channel is not. Follow one visitor from first click to paid subscription, then copy the UTM template and one-page report layout.

Most teams can see their traffic and their signups. Far fewer can answer the question the budget meeting actually asks: which channel produced customers who pay, and how much do they pay?
Getting marketing attribution to revenue means following one person from a first visit to a lead, a deal and a paying subscription, without losing the thread at any step. Below: the four stages, where the thread breaks, a minimal tracking setup, a UTM naming template and a one-page report layout.
The four stages: visit, lead, deal, revenue
Each stage has to capture something, and hand a key to the next one. If any handoff drops the key, revenue arrives with no source attached.
Stage | What happens | What you must capture | What carries it forward |
|---|---|---|---|
Visit | Someone lands on a page | Source, medium, campaign, ad click-id, landing page | A first-party visitor and session record |
Lead | They fill in a form or sign up | Contact details, plus the session's source fields | The contact record, linked to their sessions |
Deal | Sales or a trial turns them into an opportunity | Deal value, stage, close date | The deal, linked to the contact |
Revenue | They pay, upgrade, downgrade or cancel | Plan, amount, start date, changes over time | The subscription or invoice, linked to the deal or account |
A worked example (illustrative)
Take a reader of this blog. The details below are illustrative, but the path is typical.
Visit. They click a link to this post in a newsletter. The link carries
utm_source=newsletter,utm_medium=emailandutm_campaign=2026-11-attribution-guide. A session is recorded with those values.Return visit. Three days later they come back by typing the address. That session has no campaign tags and shows up as direct.
Lead. On the second visit they start a trial through the signup form. If the form carries the visitor's earlier session history, the new contact is tied to the newsletter visit as well as the direct one.
Deal and revenue. The trial converts to a paid monthly subscription. The subscription is linked to the account, the account to the contact, and the contact to both sessions.
Did the newsletter or the direct visit produce that customer? It depends on the attribution model you choose: credit to the first touch, the last touch or a split. There is no universally right answer. What matters is that you pick one, write it down, and keep using it so month-to-month comparisons mean something.

Where attribution breaks
Separate tools. Analytics knows the campaign but not the deal; the CRM knows the deal but not the campaign; billing knows the revenue but neither of the others. Joining them takes exports and a shared key that rarely survives intact. (We will put a number on that kind of tool sprawl in a later post on what a stitched revenue stack really costs.)
Missing or inconsistent UTMs. "Email", "email" and "e-mail" become three channels. Untagged social posts and partner links land in direct.
Anonymous sessions that never join a contact. If the form or signup does not pick up the visitor's session, the first visit is lost even though it happened.
Redirects and domain hops. A redirect or a move from the marketing site to the app can strip query strings, so the signup step never sees the campaign.
Revenue that changes after the deal. Upgrades, downgrades and churn happen later. If they are not tied back to the original account, a channel looks better or worse than it is.
A minimal tracking setup
You need seven things, done consistently.
Install one first-party tracking snippet on every page, including the blog, landing pages and the signup flow.
Adopt a UTM convention (template below) and tag every link you control.
Capture UTM parameters and ad click-ids on landing, and carry them into every form submission.
Create the contact at the form, and link it to the visitor's prior sessions.
Link every deal to a contact, and every subscription or invoice to that deal's account.
Connect ad spend, so return on ad spend is calculated from closed revenue rather than estimated.
Write down your attribution model and reporting window.
This is what Vertex marketing analytics is built to do. The Vertex Pixel is a first-party snippet, installed with one script tag and using no third-party cookies. It captures visitors, sessions, UTMs and ad click-ids (gclid, fbclid and more), and ties them to the contacts and deals already in your CRM. Forms submitted through Vertex inherit the UTM parameters captured on landing. Once you link your Google and Meta ad accounts through integrations, return on ad spend is computed per campaign, channel and country.
UTM naming template
Copy this into your team's link-building sheet. The rules matter more than the exact values: lowercase, hyphens instead of spaces, and a fixed list for medium.
Parameter | What it records | Rule | Example |
|---|---|---|---|
utm_source | Where the click came from | Lowercase, one name per source, never changes | newsletter, partner-blog, search-ads |
utm_medium | The channel type | Pick only from a fixed list: email, paid-search, paid-social, organic-social, referral, partner | |
utm_campaign | The campaign or launch | yyyy-mm-short-name | 2026-11-attribution-guide |
utm_content | Which link or creative | Placement or variant | header-link, cta-button, variant-b |
utm_term | Paid search keyword (optional) | Only for paid search | revenue-attribution |
Three more rules keep the data clean:
Never tag internal links between your own pages. It can overwrite the visitor's original source mid-visit.
Log every tagged link you build in one shared sheet, with the owner and date.
Review the source and medium values once a month and merge any strays.

How to read a channel-to-revenue report
Fit the report on one page and read it from right to left: revenue first, traffic last.
Channel (utm_medium) | Visitors | Leads | Visit to lead % | Deals or trials | Won | New MRR | Revenue per visitor | Ad spend | ROAS |
|---|---|---|---|---|---|---|---|---|---|
paid-search | |||||||||
paid-social | |||||||||
organic-social | |||||||||
referral and partner | |||||||||
direct and untagged | |||||||||
Total |
How to read it:
Start with revenue per visitor, not volume. A small channel that converts to paying customers can beat a large one that produces signups who never pay.
Allow for lag. Trials, sales cycles and billing dates mean this month's visitors pay next month or later. Compare periods long enough to cover your usual time to first payment.
Watch the direct row. If it is large, the problem is usually tagging, not word of mouth.
Spot-check the path. Pick a few won customers per channel and follow each back to their sessions. In Vertex you can drill any funnel stage down to a single session and see the pages, the source and the path to conversion. If the path looks implausible, fix the tracking before trusting the totals.
What not to conclude from small samples
Channel reports invite strong conclusions from thin data. Before you move budget:
Count the won deals, not just the percentages. Three wins against one is not a trend. As a rule of thumb (our judgement, not a benchmark), wait for a few dozen wins per channel before reallocating spend.
Check for one large customer. A single big account can make a channel look like a winner. Look at the count and the median alongside the total.
Do not compare across a model change. Switching from first-touch to last-touch credit moves revenue between channels without changing what happened.
Do not read a recent month as final. Late-converting trials will still land there.
Put it together
Pick the model, install the snippet, enforce the UTM template and fill in the report for last quarter. Even a partial version shows where your tracking breaks. Vertex keeps the full chain, from visitor to paying subscriber, on one database, so the report is a view rather than a monthly rebuild.
See which channel pays
Follow every visitor from first session to paying subscriber on one database, with a 14-day trial and no credit card.