What is A/B testing?
Learn what A/B testing means, how a controlled marketing experiment works and which mistakes make test results unreliable.
Marketing glossary
UTM parameters help analytics software distinguish visits from particular links and campaigns. A small naming convention agreed in advance prevents a surprisingly large future archaeology project inside your reports.

The parameters appear after a question mark in a URL. The five standard labels are:
linkedin or newsletter;paid-social, email or qr;Source, medium and campaign are the usual minimum. Term and content are optional when they answer a real reporting question.
A consultancy promotes a webinar through its email newsletter, LinkedIn adverts and a printed letter with a QR code. It creates separate tagged links:
?utm_source=newsletter&utm_medium=email&utm_campaign=ai-webinar-sep-2026
?utm_source=linkedin&utm_medium=paid-social&utm_campaign=ai-webinar-sep-2026
?utm_source=direct-mail&utm_medium=qr&utm_campaign=ai-webinar-sep-2026
The destination remains the same, while analytics can group the shared campaign and compare recorded sources. The consultancy should test each link, keep names lower-case and store them in a campaign register rather than allowing everyone to improvise variants.
Consistent tagging can:
utm_content is useful;They are particularly helpful for channels that send people through links you control, including QR codes, partner newsletters, email signatures and downloadable documents.
Inconsistent naming fragments one campaign into several rows. LinkedIn, linkedin and linked-in may be reported separately. Establish controlled values for common sources and media, then generate URLs from those values.
Do not tag internal website links. Clicking a UTM-tagged internal link can overwrite the visitor’s original acquisition information and create a false new source. Use event tracking or ordinary navigation analysis instead.
Avoid placing personal data or confidential information in parameters. URLs can appear in browser history, analytics, server logs and referrer information. A campaign label should describe marketing activity, not contain an individual’s email address or sensitive characteristics.
Finally, UTMs do not solve every attribution gap. Links may be copied, parameters stripped, cookies refused and journeys continued on another device or by telephone. They improve observation; they do not create a complete causal history.
They normally load the same page while passing tracking labels in the URL. The website and analytics configuration must preserve and record them correctly, so test tagged links before publishing a campaign.
Use a documented, consistent convention, commonly lower-case words separated by hyphens. Spaces become encoded and capitalisation can split reports, so paid-social is usually easier to manage than improvised variations.
Who wrote this
Read next
Learn what A/B testing means, how a controlled marketing experiment works and which mistakes make test results unreliable.
Learn what answer engine optimisation means and how to make useful business content easier for search and AI systems to extract and cite.
Learn what bounce rate means, why analytics tools calculate it differently and how to diagnose a high bounce rate without guessing.
A short discovery call is usually enough to tell. No deck, no pitch.
Book a discovery call