Measurement guide 04
Add UTM tags (short labels on the end of a link) to emails, flyers, and other outside links so GA4 can tell which campaign brought each visit. Use one agreed naming sheet, test every link before launch, and say which GA4 report you are reading. Then check the campaign against real inquiries before moving budget.
← All measurement and analytics guidanceAgree names
Test links
Pick the report
Check inquiries
Decide carefully
Why this matters
An owner asking whether an email worked needs more than a traffic spike. Did the right people arrive, could they contact you, did the inquiries reach the team, and did they become jobs? A campaign label helps connect those steps. It cannot capture a word-of-mouth referral or prove what would have happened without the campaign.
Tagging is a team habit. If each person invents names, one promotion splits into several report rows. If the question changes halfway through, the same numbers tell conflicting stories. A short naming sheet and a written report definition often fix more than a new dashboard. Leave uncertain results uncertain rather than forcing every inquiry into a channel.
Responsible practices
UTM (Urchin Tracking Module) tags are labels added to a link. Use utm_source for who sent it, utm_medium for the channel, and utm_campaign for the promotion, all in lowercase because values are case-sensitive. Hypothetical example: https://example.com/consultation/?utm_source=newsletter&utm_medium=email&utm_campaign=autumn_offer. Record each finished link, its owner, and launch date.
Google Ads can tag its own links automatically, which gives GA4 richer ad data. Check that link and auto-tagging before adding your own tags to ads. Auto-tagged values normally take priority. If another tool needs UTMs on ads too, test that both sets of values survive.
Do not add UTMs to your own menus, banners, or links between pages; that restarts the visit's source. Track those clicks with an event instead. Keep names, emails, and phone numbers out of tags and links. Label the placement, never the customer, and check the full link before sending it.
First user source shows how someone first found you. Session source shows what brought each visit. GA4 can also give credit for a key event to the channel that brought the visit. By default it uses a 'data-driven' model; the other option is last click. Name the report in the heading, and never add these totals together.
Open a sample from the real email, QR code, or ad, including any link shortener. Confirm the right page loads, the tags survive, and GA4 sees the visit. If a campaign seems missing, check the saved link, redirects, setup, date range, and report, in that order. Keep 'direct' and 'unassigned' visible rather than reassigning them.
Hypothetical example: campaign A brings 120 tagged sessions and six confirmed inquiries; campaign B brings 60 sessions and four. Their inquiry rates are 5% and about 6.7%. That alone does not make B better: the counts are small, costs may differ, and quality is unknown. Check cost and fit before scaling either one.
Campaign tracking
Write one question for the campaign, such as whether a newsletter brings consultation requests.
Add the source, medium, campaign, link, owner, and launch date to the naming sheet.
Have a second person check each finished link for spelling and personal information.
List every place the link will appear, and note where Google Ads auto-tagging already applies.
Test a link from each placement on a phone and a computer, including the contact path.
Save the report settings you will use: date range, time zone, filters, and source report.
Review tagged visits against confirmed inquiries, then record one next step, an owner, and a date.
Avoid these failures
Collection, attribution, qualification, bookings, costs, and revenue depend on technical, customer, operational, and platform factors outside any provider's control. Results are not guaranteed.
Questions business owners ask
No. Tag outside marketing links where a separate label answers a real question. Leave links within your own site untagged. Skip tags for minor variations no one will act on.
No. The model affects which channel gets credit for key events. First-user and session source reports are not changed by it. Note the model with each report, and do not present a change in credit as new customers.
Not always. It also includes visits where the source was lost, such as links in some apps or untagged documents. Check campaign links if its share changes suddenly, and leave the rest labeled as unknown.
No. They show that recorded visits and outcomes happened together. Proving cause needs a careful test, such as pausing a campaign in one area. For everyday decisions, note other explanations and make small, reviewable changes.
Primary guidance
Give each campaign a traceable starting point