utm_source vs utm_medium: the rule that fixes most GA4 attribution

The single mistake that creates more messy GA4 reports than any other is putting the channel name in utm_medium. utm_medium=facebook instead of utm_medium=social. Every time someone makes that mistake, their Facebook traffic vanishes from the Paid Social row and lands in "Other," which is the GA4 graveyard for misconfigured tags.

Here is the rule, the lookup table that fixes it, and the half-dozen edge cases that come up.

The rule

utm_source is the specific origin of the click. utm_medium is the channel category that origin falls into. One source, one medium, every time.

• utm_source=facebook and utm_medium=social (Facebook organic post)
• utm_source=facebook and utm_medium=paid_social (Facebook ad)
• utm_source=mailchimp and utm_medium=email (any email through Mailchimp)
• utm_source=linkedin and utm_medium=social (LinkedIn organic post)
• utm_source=linkedin and utm_medium=paid_social (LinkedIn ad)
• utm_source=google and utm_medium=cpc (Google paid search)

The source is always specific. The medium is always categorical.

Why GA4 cares about utm_medium specifically

GA4's default channel grouping is built around utm_medium. The exact strings GA4 looks for, in the channels most marketing teams use:

GA4 channel group	utm_medium values it matches
Paid Search	cpc, ppc, paidsearch
Paid Social	paidsocial, paid_social, social-paid, paid-social
Organic Social	social, social-network, social-media, sm
Email	email, e-mail, e_mail
Display	display, banner, expandable, cpm
Affiliates	affiliate, affiliates
Referral	referral
Organic Shopping	organic_shopping
Paid Shopping	paid_shopping, paidshopping

Any value that does not match these strings goes into "Unassigned" or "Other" depending on the GA4 version. A utm_medium=facebook lands in Other. A utm_medium=paidsocial lands cleanly in Paid Social. The tag is a one-character difference and the report is a different row.

The lookup table that keeps your team aligned

Pin this somewhere. The table below covers roughly 95 percent of marketing channels:

Channel	utm_source examples	utm_medium
Google paid search	google	cpc
Google organic	google	organic
Bing paid search	bing	cpc
Facebook ads	facebook	paid_social
Facebook organic	facebook	social
Instagram ads	instagram	paid_social
Instagram organic / bio	instagram	social
LinkedIn ads	linkedin	paid_social
LinkedIn organic	linkedin	social
TikTok ads	tiktok	paid_social
Twitter / X organic	twitter	social
Reddit organic	reddit	social
Reddit ads	reddit	paid_social
YouTube organic	youtube	social
Pinterest	pinterest	social
Mailchimp	mailchimp	email
Iterable, Klaviyo, Customer.io	(vendor name)	email
Customer newsletter	newsletter	email
Sales rep email	sales-email	email
Partner blog post	partner-name	referral
Affiliate network	(network or affiliate ID)	affiliate
Display ad network	google-display, meta-audience	display
Podcast ad	podcast-name	podcast
QR scan from print	qr	offline
SMS	sms	sms

Two notes:

1. Use lowercase, hyphens, and underscores. GA4 distinguishes paid_social and paid-social. Both appear in the GA4 channel match list. Pick one and stay consistent. Underscore is more common in GA4 documentation; hyphen reads better in URLs. We default to underscore for the medium values listed above and hyphen for source values.
2. Use cpc for paid search, paid_social for paid social. This is a GA4 oddity. Paid search uses the historic Google Ads convention cpc. Paid social uses the modern paid_social. Both work; mixing them inside one campaign does not.

The three edge cases that trip teams up

Influencer posts. If a creator posts about your product on their channel, the source is the creator (utm_source=creator-name) and the medium depends on how you compensated them. Paid sponsorship is paid_social. Affiliate commission is affiliate. Earned-media organic mention is referral or social depending on the channel.

Email links inside email. A weekly newsletter contains links. A subscriber clicks. Source is the newsletter platform (mailchimp, klaviyo). Medium is email. So far so good. The trap: when the visitor lands and clicks an internal link on your site, GA4 attributes a new pageview to the original source as long as the session continues. This is correct. The mistake some teams make is appending UTMs to internal site links inside the same session, which breaks the attribution chain.

Cross-domain campaigns. Your campaign sends a visitor from marketing.com to app.com. Without cross-domain measurement enabled in GA4, the second domain shows up as a referrer and the original utm_source and utm_medium get clobbered. Fix: Admin → Data Streams → Configure tag settings → Configure your domains. Add both domains.

The smart-paste shortcut

Manually typing source-medium pairs into a UTM builder is the slowest path to the right answer. The faster path:

1. Build the URL once for each channel using a builder that enforces the table above.
2. Save the templates in a shared doc your team works from.
3. When a new campaign starts, copy the template URL and replace only the utm_campaign value.

The Trakl free UTM builder has the source/medium taxonomy from the table above built into dropdowns. The medium dropdown filters to whatever is valid for the source you picked, so a marketer cannot pick linkedin / cpc even by mistake.

What to do if your historical data is already a mess

Three steps:

1. Open GA4. Reports → Acquisition → Traffic Acquisition. Group by Source / Medium.
2. Sort by sessions descending. Look at the top 30 rows.
3. For every row where medium is not in the GA4 channel list, write down the campaign that produced it.

Now you have a hit list. Going forward, every link your team builds gets the right utm_medium from the start. Old data stays wrong; you cannot fix the past without a republished tracking plan, which is rarely worth it. The point is to stop the bleeding.

For the longer rules around naming, casing, and the team-discipline question, the piece on utm naming conventions is the next read.