Audit your UTM tags and tracking setup

Kite begins by collecting the tag values your tools actually received, because the real state of your tagging lives in your reports, not in anyone's memory. In Google Analytics, it runs reports of sessions and conversions by source, medium, and campaign. In PostHog, it runs PostHog's own UTM audit, which finds campaigns with mistyped or unmatched sources. In Mixpanel, it reads the values stored for each UTM property. On a site Kite hosts, it reads its own built-in analytics.

Then Kite groups the values that mean the same thing. facebook, Facebook, fb, and facebook.com become one source. cpc, CPC, and paid-search become one medium. Campaign names get the same treatment, so spring_launch, Spring Launch, and springlaunch2026 show up as one campaign that was tagged three ways. For each group, Kite records how many visits and conversions sit under each spelling, so you can see how much each inconsistency hides.

If your team already keeps a spreadsheet of campaign links or a written convention, share it with Kite. It compares what the convention says with what people actually did, which usually shows exactly where the drift started.

Kite checks which channel each group of tags ends up in, because a tag can be spelled consistently and still land in the wrong place. Google Analytics assigns every visit to a channel group, such as Paid Social, Email, or Organic Search, by matching the source and medium against its rules. Kite reads the channel groups defined on your property and the share of visits and conversions that fall into Unassigned. When a medium like newsletter, social-paid, or banner matches no rule, Kite names it and shows which campaigns used it.

Direct traffic gets the same scrutiny. Kite looks at which pages receive the most direct visits. A homepage with plenty of direct traffic is normal, because people type your address. A pricing page or a campaign landing page with mostly direct traffic usually means the links pointing to it went out without tags, often in emails, PDFs, or messages shared by hand.

Kite's own analytics works the same way on sites it hosts. It sorts each visit into paid search, paid social, email, AI assistants, organic search, organic social, referral, or direct, based on the link's tags and the site the visitor came from. It recognizes mediums such as cpc, paid_social, and email, so an email tagged newsletter with no referring site gets counted as direct there too.

Kite tests the tracking itself, because a perfectly tagged link is worthless if the visit never gets recorded with its tags intact. It opens a sample of your tagged links in a browser and checks that the tags are still on the address when the page finishes loading. A redirect from an old campaign address, a move from http to https, or a link shortener can drop everything after the question mark, which turns a tagged click into a direct visit.

Next, Kite renders your homepage, your campaign landing pages, and the pages along your signup or checkout path, and it reads which analytics tags each one loads. A page that loads no tag records nothing. A page that loads Google's tag twice, once directly and once through Google Tag Manager, doubles its page views. A page that reports to a different Google Analytics property sends its data somewhere your team isn't looking.

Kite also looks for the places where a visitor's original source gets replaced. When your own domain, your app's domain, or a payment page such as checkout.stripe.com shows up as a source of conversions, the credit for those conversions was lost at that step. The same thing happens when your own website uses UTM tags on links between its own pages, because each internal click overwrites the campaign that brought the visitor in. Kite lists every internal link it finds carrying tags.

When your events flow through Segment, Kite checks whether they arrive where they should. It reads each source's daily event counts and each destination's delivery numbers, so a drop that starts on the day of a release points to the release rather than to your campaigns.

A good UTM naming convention uses lowercase letters, one separator, a short fixed list of mediums your analytics tool already recognizes, and campaign names built from the same parts in the same order every time. Kite writes that convention for your team from the audit, because the right list of sources and campaigns depends on the channels you actually use.

A typical convention from Kite covers each tag like this:

• Source names the specific site or sender, such as linkedin, google, mailchimp, or partner-acme. Kite picks one spelling for each source you use and lists it.
• Medium names the kind of traffic, chosen from the values your analytics tool sorts into the right channel, such as cpc for paid search, paid_social for paid social, social for unpaid posts, email for email, and referral for partner links.
• Campaign names the effort, built from the same parts in the same order, such as 2026-10-webinar-ai-visibility. The date at the front keeps campaigns in order when you sort a report.
• Content tells apart the versions inside one campaign, such as two ad images or the header and footer links in one email.
• Term holds the paid search keyword when you use one.

The convention also covers the rules that keep reports clean. Everything is lowercase, because most analytics tools treat LinkedIn and linkedin as different values. Words are joined with hyphens rather than spaces. Tags go only on links that point to your site from somewhere else, never on links between your own pages. Personal details, such as a recipient's email address, stay out of tags entirely.

If your team already has a convention that mostly works, Kite keeps it and fills the gaps rather than starting over. Either way, Kite saves the final version to your team's knowledge base, so every person and every later piece of Kite's work uses the same reference.

Kite fixes the problems it can change for you and gives exact steps for the ones that live in another tool's settings. On a site Kite hosts, or one it works on through your GitHub repository, it can add a missing analytics tag, remove a duplicate one, fix a redirect that strips tags, and take UTM tags off your internal links. It shows you the changes first, and they go live when you approve them.

Most analytics tools don't let anyone rewrite the tags on visits they've already recorded, so Kite makes the existing data readable instead. In PostHog, it suggests the mappings that put mistyped sources back under the right channel. In Mixpanel, it can build a lookup table or a formula-based property that groups the spelling variants under one name, so a breakdown by source reads the way your team talks about it. In Google Analytics, it tells you which domains to add to the unwanted referrals list and how to set up cross-domain measurement when your site and app live on different domains.

Kite ranks every finding by how much it distorts the numbers you make decisions with. A payment page taking credit for every purchase comes first, because it hides which campaigns produce revenue. A source spelled two ways on a small campaign comes later.

Kite uses the convention on every campaign link it creates, so the cleanup doesn't decay the moment the audit ends. When it builds a landing page campaign, an outbound email, or a set of social posts, each link carries the agreed source, medium, and campaign, plus a content tag for each message or version. It records the campaign and its tags in your knowledge base, which gives your team a running list of what was tagged and when.

Links your team creates by hand still need checking. Ask Kite to repeat the audit every month, and each run compares the new tag values with the convention and with the last audit. A new spelling, a medium that slips into Unassigned, or a landing page that suddenly stops recording visits shows up within weeks. On a site Kite hosts, conversion tracking is added automatically every time you publish, so new pages start measuring from the day they go live.