UTM Parameters & Campaign Tracking

What UTMs are for

UTM parameters (Urchin Tracking Module — named after the analytics company Google acquired in 2005) are query string parameters appended to a URL that tell an analytics tool where a click came from. They exist because a raw referrer header only tells you the *domain* a visitor arrived from (e.g. instagram.com), not which specific post, ad, or email drove the click. UTMs let you attach structured campaign metadata to any link you control, so that metadata survives the click and lands in your analytics tool's traffic-source dimensions.

The five parameters

There are five standard UTM parameters, and only two are strictly required for GA4 to register a distinct traffic source: utm_source and utm_medium.

- utm_source — the referring platform or publisher (google, newsletter, facebook, partner-blog). Answers 'where did this click come from.'
- utm_medium — the marketing channel category (cpc, email, social, referral, banner). Answers 'what kind of channel was it.'
- utm_campaign — the specific initiative name (fall_sale_2026, q3_webinar_series). Answers 'which effort was this part of.'
- utm_content — differentiates between variants pointing at the same campaign (headline_a vs headline_b, or sidebar_cta vs footer_cta). Used for A/B testing creative within a campaign.
- utm_term — historically used for paid search keyword tracking (utm_term=martech+consultant); largely legacy with the shift to broad match and automated bidding, but still used for manually tagging audience segments in some paid social setups.

utm_example.txt

text

https://sadullah.online/services?
  utm_source=linkedin
  &utm_medium=social
  &utm_campaign=q3_martech_launch
  &utm_content=carousel_post_v2
  &utm_term=

utm_source and utm_medium are the minimum

GA4 (and most analytics tools) will still register a session-level traffic source with only utm_source and utm_medium present. utm_campaign, utm_content and utm_term are optional enrichment — but skipping them means you lose the ability to compare creative variants or initiatives later, and you can't retroactively add them to historical clicks.

Naming convention governance — the part everyone skips

UTMs are free-text strings, which means without an enforced convention, the same channel fragments into a dozen inconsistent values: Facebook, facebook, FB, fb_ads, Facebook Ads all show up as different rows in a traffic-source report, silently splitting what should be one channel's data across multiple lines. This is the single most common reason marketing analytics data looks unreliable — not a tracking failure, but a governance failure.

The fix is a written naming convention plus enforcement, not just documentation. Effective governance means: always lowercase, a fixed enumerated list of allowed utm_source and utm_medium values (not freeform), a consistent campaign naming pattern (e.g. {quarter}_{initiative}_{audience}), and a shared campaign URL builder (a Google Sheet with dropdowns, or a tool like UTM.io) that every team member uses instead of hand-typing URLs.

Common UTM naming mistakes

Mistake	Why it breaks reporting
Inconsistent casing (Facebook vs facebook)	Most tools treat UTM values as case-sensitive strings — creates duplicate rows
Freeform utm_medium values	'paid', 'cpc', 'ppc', 'paid-search' fragment what should be one channel
Tagging internal links	Tagging your own site's internal nav links overwrites the original session source, corrupting attribution
Tagging owned email footer links to your own site	Same self-referral problem — resets attribution mid-session
No utm_content on multivariate creative	Impossible to tell which ad variant actually drove the conversion
Manually typed URLs with no builder tool	Typos (utm_source=linkedIn vs linkedin) silently split data

How UTMs flow into GA4's traffic_source

When a tagged URL is clicked, GA4's tag reads the query string parameters on page load and writes them into the session's session_source, session_medium, and session_campaign fields as part of that session's traffic-source data, which is also exported to BigQuery inside the traffic_source and collected_traffic_source nested fields on each event. GA4 uses a last non-direct click model by default within its standard channel grouping reports: if a user arrives via a tagged campaign link, then returns later by typing the URL directly, GA4 still credits the original tagged session as the acquiring source for that user, rather than overwriting it with '(direct)/(none)'. This is why 'direct' traffic often looks smaller in GA4 than the raw number of direct visits would suggest.

utm_source_medium_query.sql

sql

SELECT
  (SELECT value.string_value FROM UNNEST(event_params) WHERE key = 'source') AS utm_source,
  (SELECT value.string_value FROM UNNEST(event_params) WHERE key = 'medium') AS utm_medium,
  (SELECT value.string_value FROM UNNEST(event_params) WHERE key = 'campaign') AS utm_campaign,
  COUNT(DISTINCT user_pseudo_id) AS users
FROM `project.analytics_123456789.events_*`
WHERE event_name = 'session_start'
  AND _TABLE_SUFFIX BETWEEN '20260801' AND '20260831'
GROUP BY 1, 2, 3
ORDER BY users DESC

GA4's Default Channel Grouping can override sloppy UTMs

GA4 doesn't take utm_medium at face value — it maps source/medium combinations into its own Default Channel Group using internal rules (e.g. medium containing 'cpc' or 'ppc' maps to Paid Search). If your utm_medium values don't match GA4's expected patterns, traffic can land in 'Unassigned' or the wrong channel group even though the raw UTM data was captured correctly.

What's next

Clean UTM data is the raw material every downstream analysis depends on — including A/B tests measuring which campaign variant actually converts better. Understanding how to read those results correctly is the next skill.

Next: A/B Testing & Statistical Significance →