How-To Guide
Do UTMs on internal links inflate conversions in GA4?
See how internal UTMs overwrite session attribution in GA4, which reports look wrong, and how to measure the same click with an event instead.
No. Google Analytics 4 does not restart a session when new campaign parameters are met mid-visit, so an internal UTM cannot add sessions, and a key event that fires later is still one event. Google's comparison of GA4 and Universal Analytics says sessions are not restarted when new campaign parameters are encountered, while the Universal Analytics column says the opposite rule applied there. The real cost in GA4 is labeling: campaign values from the tagged URL can attach to events after the click, which changes which campaign or channel gets credit and can leave Unassigned, (not set) or Direct rows that look like inflation. Remove the tag from internal links and measure the interaction as an event.
- Sessions do not increase, because the session keeps the source it started with; a 30-minute inactivity gap is what ends it.
- Event-scoped campaign fields can take the internal label, so a report grouped by campaign can disagree with the same report grouped by session.
- Channel rules turn a partial or unknown label into Unassigned, (not set) or Direct, which is the row most people read as inflated traffic.
The short answer
No. In Google Analytics 4 an internal link carrying UTM parameters does not start a new session, so it cannot multiply sessions, and a key event that fires later is still one event. Google's comparison of GA4 and Universal Analytics states the rule in the GA4 column: "Sessions aren't restarted at midnight or when new campaign parameters are encountered." The same page records the older rule for Universal Analytics, where picking up new campaign parameters mid-visit did start a new session.
That difference is why the question keeps getting a yes. Many guides that warn about inflated sessions were written for Universal Analytics, or repeat a paragraph written for it. The tag is still a poor choice in GA4, but the damage is a labeling problem: a report that credits the wrong campaign, not a count that grows.
- The session keeps the traffic source it started with, so session counts do not rise.
- Campaign values from the tagged URL can attach to events collected after the click, which splits where the story is read.
- Channel rules then file the row somewhere you did not choose: Unassigned, (not set) or Direct.
The rest of this page separates the two effects, shows how to tell them apart in your own property, and gives the replacement measurement for an internal link.
Original diagram of the session rule in both products. The rule text is from Google's GA4 versus Universal Analytics comparison page, read 16 November 2026.
Make it work for your brand
Find it. Remake it. Post it. Superdirector turns the viral formats in your niche into your brand's carousels, X posts and video scripts, and keeps the calendar full.
Why the inflated-session story comes from Universal Analytics
The session rule changed between the two products, and the change is in the documentation. Read the four rows together and the old advice makes sense for the old tool.
| Behaviour | Universal Analytics | Google Analytics 4 | Source |
|---|---|---|---|
| New campaign parameters met mid-visit | Starts a new session | Session continues | Google comparison doc |
| A brand tagging links on its own site | Produces a much higher session count | Lower session count than UA | Google comparison doc |
| Where the tag's values apply | Session source and medium | Events collected after the click | LinkMngr |
| What the reader sees | Extra visits in the session count | A row carrying the wrong campaign label | This article's reading of the sources |
Google is explicit that the gap widens when a site tags its own links: "If you do use UTMs on your own website, you may notice a much higher count of sessions in UA than in Google Analytics." The direction of that gap is the clearest signal that the inflation story belongs to the older product. The same guidance still says not to tag your own site. Its stated reason is the session reset in Universal Analytics; in GA4 the cost has moved from the session count to the labels.
What an internal UTM changes in Google Analytics 4
The session survives, and the session-scoped fields keep their entry values. What changes is the event scope. LinkMngr's worked example describes the split in one sentence: "The session may still show newsletter / email while later event-scoped campaign dimensions show homepage / internal." A visitor who arrived from a newsletter can therefore appear under a navigation label on the events that follow a tagged menu click, while the session itself still belongs to the newsletter.
The reach of one parameter is wider than most people expect. Google's manual-tagging page says that when any UTM parameter is present, "Google Analytics will derive all cross-channel traffic source dimension values from UTM parameters, exclusively." A stray utm_source on a pricing link can replace the whole cross-channel reading for the events after it, not only the field someone typed.
Conversions follow the same path. A GA4 conversion is a key event, and any event you collect can be marked as one. Key event reporting attributes credit to touchpoints along the user's path, so a changed campaign label moves credit between rows. The number of key events stays the same; the row that receives the credit changes. That is why a team can swear the numbers are inflated and still find a flat total when it counts events instead of rows.
Why the report looks inflated when the counts are flat
Three mechanisms produce a wrong-looking traffic report without adding a session or a conversion.
- Channel rules. Channels are rule-based, and a source value that matches no rule lands in Unassigned. A tag with a missing campaign field leaves (not set) in the campaign column instead.
- Inheritance. A later direct session can carry the previous session's UTM value. The lookback window for a session is set by the Campaign timeout, which is six months by default, so one stray label can follow a visitor for a long time.
- Duplicates and redirects. The same key can appear twice in a query string, and a redirect can strip the tag on the way. One failure adds a conflicting label, the other sends the visit toward Direct.
The official wording is plain. "Analytics categorizes traffic sources into channels based on fixed rules," and "the unassigned row appears when Analytics cannot categorize the source of traffic." Neither line mentions sessions, because neither problem is a session problem.
Duplicates deserve their own warning because they break the habit of trusting the URL you can see. UTMCraft's duplicate-key guide defines a campaign URL as duplicated when "the same tracking key appears more than once in its query string," and adds that "There is no reliable universal rule that the first or last occurrence wins across every consumer." Two teams that each added a source value can both be right about their own link and wrong about the one that shipped.
Inheritance is the effect most often misread as inflation. Analytics Mania's UTM guide names the mechanism: if the next session is direct, the UTM value from the previous session will be inherited, and the page says the result distorts report accuracy. The session count did not move; the row that owns the visit did.
Original diagram of where a mislabelled campaign value shows up. Row definitions are from Google Analytics Help, read 16 November 2026.
How to check your own property
The test needs no new tool and no code change. Run it before you change any live link, so the before-and-after comparison is honest.
- Open Reports, then Acquisition, then Traffic acquisition. Put Session source / medium beside Session campaign and read one row at a time.
- Add the event-scoped campaign dimension from Explore. If the session scope and the event scope disagree on the same visitor, a mid-session tag is the likely cause.
- Filter to the label you suspect, such as internal, homepage or nav. Compare sessions and key events by date, not as a single total.
- Search the site for internal URLs containing utm_. A site search or a crawl finds menu, footer and landing-page links faster than a manual browse.
- Inspect one suspect URL for a repeated utm_source key and for a redirect that drops the query string on the way to the final page.
- Fix the template or the CMS field that produced the link, then re-read the same two reports after a week before you delete anything else.
How to remove the tags without losing the measurement
Removing the tag is the easy half. The team added it because someone wanted to know which menu item or banner earned the click, and that question is legitimate. Measure the interaction as an event instead of an arrival.
- Navigation and editorial links: send an internal_link_click event with the link URL, the visible text and the page area as parameters, and leave the href clean.
- Promotion banners: use the recommended view_promotion and select_promotion events so the banner is measured without inventing an acquisition source.
- Links that leave the site: let Enhanced Measurement collect the outbound click event, which keeps the destination session separate.
- Owned domains in one journey: configure cross-domain measurement instead of copying a source label onto the link.
Keep UTM parameters for the job they were documented for. Google describes them as parameters on the destination URLs used in referral links and ad campaigns, so an external campaign link keeps its tags. Inside that external set, keep one unique source value per platform, since a standardized strategy is what keeps attribution accurate without fragmenting data.
Original mapping of link type to event. Event names come from the practitioner sources named in the article, read 16 November 2026.
When a tagged internal URL is legitimate
The honest exceptions are narrow. A link that genuinely crosses a measurement boundary can need a label: an owned domain that is not yet configured for cross-domain measurement, a landing page handed to a partner system, or a paid placement whose destination is your own page and whose platform requires campaign parameters. In each case the label describes an arrival from outside, which is exactly what the parameter was designed to record.
A menu item, a related-post card, a footer link or a banner between two pages of one property is not that case. Those interactions belong in events. When a partner or an agency asks for a tagged internal link, ask which boundary it crosses; if the answer is none, the event is the measurement they actually want.
Where Superdirector fits
Clean attribution matters most when the report drives a decision about the account itself. Superdirector's Growth plan at $199 a month runs one AI-operated social account: it researches what is working in the niche, builds carousels and X posts, schedules them and publishes to supported Instagram and X accounts under account-level standing posting permission. The Starter plan at $29 a month is the Claude or ChatGPT lane, where the same data, ideas, scripts and downloads are available but the account is not run for you.
The measurement advice on this page applies either way, because a mislabelled campaign row is misleading whether a person or a system reads it next.