UTM Parameters

What are UTM parameters?

UTM parameters are small pieces of text appended to a URL after a question mark that tell analytics tools where a click came from. The acronym stands for Urchin Tracking Module, named after Urchin Software Corp, the web analytics company Google acquired in 2005 and rebuilt into Google Analytics. The tags were originally Urchin's convention for cross-channel campaign tracking; Google kept the syntax when it folded Urchin into the new Google Analytics product, and the convention has been the de facto industry standard for two decades.

A tagged URL looks like this: https://example.com/landing?utm_source=chatgpt&utm_medium=cpc&utm_campaign=q2-launch. The question mark separates the page path from the parameters; the ampersand separates one parameter from the next. When a visitor clicks, the tags travel with them into their browser, then into the analytics tool that reads the URL on page load, then into a permanent label on that session in the analytics database.

UTM parameters are distinct from auto-tagging, which is what ad platforms do automatically with their own click identifiers (Google Ads gclid, Microsoft Ads msclkid, Meta fbclid). Auto-tagging is platform-specific and captures granular click-level data; UTMs are platform-agnostic and capture campaign-level attribution that every analytics tool can read. Most teams use both.

The 5 standard UTM parameters

Five tags are universally recognized. Three are required for any clean attribution setup; two are optional.

utm_source (required) - the channel

The platform or origin of the traffic. Values are typically short channel names: chatgpt, google, linkedin, newsletter, partner-site. One value per channel; consistency across campaigns is what makes channel-level reporting possible.

utm_medium (required) - the type of traffic

The bucket the traffic belongs in: cpc for cost-per-click ads, email for email campaigns, social for organic social posts, referral for partner links, display for display advertising. Google Analytics 4's default channel grouping uses utm_medium values to auto-classify traffic into named channels (Paid Search, Email, Organic Social, Referral). Pick values that match GA4's defaults if you want automatic classification.

utm_campaign (required) - the campaign name

The specific campaign the click belongs to: q2-launch, black-friday-2026, aeo-playbook-promo. Naming conventions matter here. Kebab-case-lowercase is the durable default; UTM values are case-sensitive in most analytics tools, so Q2-Launch and q2-launch become two separate campaigns in your reports.

utm_content (optional) - the creative variant

The specific ad, link, or button within a campaign. Useful for A/B testing two creative variants of the same campaign: headline-a vs headline-b, banner-300x250 vs banner-728x90. Without utm_content, you cannot tell which creative drove which conversions.

utm_term (optional) - the keyword or audience

Originally used for paid-search keywords; today often repurposed for audience segments, context-hint clusters in ChatGPT Ads, or LinkedIn audience IDs. Optional but useful when keyword-level or audience-level reporting matters.

How UTM parameters work

The mechanics are simple. The marketer constructs a tagged URL, places it as the destination on an ad, link, email, or social post. A visitor clicks; the browser navigates to the destination; the destination page's analytics tag (Google Analytics 4, HubSpot tracking script, Salesforce Web-to-Lead form, etc.) reads the query string and records the parameter values as session-level attributes. The session is now labeled with its source, medium, and campaign. Downstream reports, dashboards, and attribution models can filter on those labels.

UTM parameters persist for the life of the session in most analytics tools and are written into the contact record in most CRMs the first time a tagged visitor converts. After the conversion, the labels become permanent on the contact and can be used for cohort analysis, channel-level ROAS calculations, or paid-channel optimization decisions.

Three platform behaviors are worth noting. Google Analytics 4 reads UTM parameters automatically and maps utm_medium values to default channels (cpc → Paid Search, email → Email, etc.). HubSpot captures UTM parameters as contact properties named hs_analytics_source and Original Source Drill-Downs 1 and 2; the first-touch UTM is stored on contact creation. Salesforce ingests UTMs through Web-to-Lead form forwarding or Pardot/Marketing Cloud Account Engagement tracking; the destination field depends on which Salesforce form layer captures the lead.

UTM parameters vs auto-tagging

Two distinct mechanisms accomplish overlapping but different goals.

Auto-tagging is when the ad platform itself appends a unique click identifier to every outbound click: gclid for Google Ads, msclkid for Microsoft Ads, fbclid for Meta, li_fat_id for LinkedIn, wbraid and gbraid for app-tracking on Google. The identifier resolves to detailed click-level data inside the ad platform's reporting and unlocks offline-conversion uploads (you can send a closed-won opportunity back into Google Ads with the gclid, and Google credits the right keyword). It does not appear in third-party analytics tools as a named channel by default.

UTM parameters are platform-agnostic and human-readable. They travel into Google Analytics 4, HubSpot, Salesforce, Mixpanel, Segment, Amplitude, or any other tool that reads URL query strings. They do not unlock platform-side optimization features (you cannot import an offline conversion against a UTM the way you can against a gclid), but they let cross-platform attribution work without per-tool integration setup.

Most mature paid-marketing programs use both: auto-tagging on the ad platform's click for bid optimization and offline-conversion import, plus UTMs on the same destination URL for cross-platform reporting. The two coexist on the URL; nothing breaks when both are present. For platforms that have not yet shipped auto-tagging (ChatGPT Ads Manager Beta is one as of May 2026), UTMs are the only attribution layer available.

UTM parameters in AI search

UTM parameters show up two different ways in the AI search era. Both use the same syntax; the responsibility for adding them differs.

Publisher-side (organic): AI platforms auto-append their own UTMs

Starting in 2025, ChatGPT began auto-appending utm_source=chatgpt.com to outbound citation links in its responses. Perplexity has experimented with similar tagging on some answer types. The parameter shows up in publishers' analytics as a clean signal that the visit came from an organic ChatGPT citation rather than a generic referral. This is the AI platform's auto-tagging on links it sends to publishers, not something the publisher controls. It partially closes the dark AI traffic blind spot for sites that previously saw AI referrals classified as (direct) / (none) in AI referral traffic reports.

Advertiser-side (paid): marketers tag their own destination URLs

Inside ChatGPT Ads Manager Beta, advertisers add UTM parameters to the destination URL of their ad the same way they would for any other paid platform. OpenAI's help center confirms: "Advertisers can also track traffic from ChatGPT Ads by adding static tracking parameters (e.g. UTM parameters) to landing page URLs. These parameters persist on ad clicks and can be used in existing analytics tools to understand traffic from ChatGPT Ads." For the operational walkthrough, see our guide on how to add UTM codes to ChatGPT Ads.

How to measure UTM-tagged traffic

Each major destination platform reads UTMs differently. Three quick orientation notes; full how-tos live in the linked guides.

Google Analytics 4

GA4 reads UTMs automatically. The utm_source and utm_medium values map to default channel groupings unless a custom channel group overrides them. Build a custom report filtered to session_source = chatgpt and session_medium = cpc to isolate ChatGPT Ads performance. UTM-driven channel reporting is one of the things the AI-Advisors GA4 integration overlays alongside AI bot traffic and AI referral data.

HubSpot

HubSpot captures UTMs as standard contact properties (hs_analytics_source, Original Source Drill-Down 1, Drill-Down 2). The first-touch UTMs are written on contact creation; subsequent visits do not overwrite them by default. Build a contact saved view filtered to Original Source Drill-Down 1 = chatgpt and pin it to the marketing dashboard.

Salesforce

Salesforce ingests UTMs through Web-to-Lead form mapping or Pardot/Marketing Cloud Account Engagement tracking. The destination field depends on which form layer captures the lead; common patterns map utm_source to LeadSource or a custom Campaign Source field. The CRM contact record becomes the source of truth for downstream pipeline-attribution work.

Common UTM mistakes

Case-sensitivity drift

utm_source=Chatgpt and utm_source=chatgpt are two different channels in most analytics tools. The cure is a written convention everyone tags against (kebab-case-lowercase is the safest default) and a URL builder that enforces it rather than ad-hoc typing.

Inconsistent values across campaigns

One campaign uses utm_source=chatgpt; the next uses utm_source=chatgpt-ads; the third uses utm_source=openai. Three campaigns end up reporting as three different channels. Lock the convention before the first campaign ships and audit it quarterly.

Encoding issues on URLs with special characters

Spaces, ampersands, and other special characters in UTM values can break the URL or get parsed inconsistently. Stick to URL-safe values (letters, numbers, hyphens, underscores). If a value has to contain a space, URL-encode it: q2%20launch, not q2 launch.

UTMs on internal links

Tagging internal links inside your own site overwrites the original source attribution when the visitor clicks through. A visitor arrives from ChatGPT Ads with utm_source=chatgpt, then clicks an internal CTA tagged with utm_source=blog. Most analytics tools now read the second UTM as the new source, and the original attribution is lost. Reserve UTMs for external links only.

Frequently asked questions