by Om RathodMeta says one thing. TikTok says another. Your GA4 dashboard says a third. If you're running ecommerce ads across more than a couple of platforms without a plan for UTM parameters for ecommerce campaigns, you don't have "slightly messy" data. You have no real way to know which channel actually drove the sale.
That's the problem UTMs solve. Not perfectly, but reliably, if you set them up right and stop letting every team member invent their own naming scheme on the fly.
Every ad platform grades its own homework. Meta's reported conversions, TikTok's reported conversions, Google's reported conversions: add them up and they'll almost always exceed your actual total revenue. Each platform wants credit for the sale, so each one takes it. Without UTMs tying a session to a specific campaign, you have no independent referee.
This got worse after iOS 14.5. Pixel-based attribution took a real hit, and cookie restrictions keep chipping away at what platforms can see on their own. UTMs are one of the few signals left that you fully control. Nobody can degrade a parameter you appended to your own URL.
Here's the blunt version: a brand running paid social, paid search, email, and affiliate without a consistent UTM structure isn't "missing some data." It's flying blind on true channel performance. You can still see clicks and spend. You just can't reliably connect either one to actual revenue.
A UTM parameter is just a tagged bit of text appended to a URL, and GA4 (or any analytics tool) reads it to bucket the session correctly. There are five:
Here's the same product page tagged for a Meta ad and a Klaviyo email:
Meta ad:https://yourstore.com/products/hoodie?utm_source=meta&utm_medium=paid-social&utm_campaign=q4_bfcm_launch&utm_content=ugc_video_v2
Klaviyo email:https://yourstore.com/products/hoodie?utm_source=klaviyo&utm_medium=email&utm_campaign=bfcm_flow_abandoned_cart
Same URL, same product, completely different reporting rows in GA4. That's the point.
For basic tracking, you need three: source, medium, campaign. Those three are what actually separate your channels in reports. Term and content are optional, but skip them and you lose the ability to see which specific ad creative or audience segment is pulling weight inside a campaign, which matters a lot once you're A/B testing.
GA4 treats capitalization and formatting literally. "Summer_Sale" and "summer-sale" and "Summer Sale" are three different campaigns as far as GA4 is concerned, even though a human reads them as identical. This is the single most common way ecommerce reporting quietly turns into a mess: nobody typos the campaign, they just format it differently.
Fix: lowercase only, underscores as separators, no spaces, ever. Full stop.
Then lock a taxonomy before anyone builds a link:
Put this in a shared doc or spreadsheet with columns for source, medium, campaign, content, and date launched. Every new link gets logged there before it goes live, not after someone notices the reporting looks weird. This is tedious in month one. It's the reason your reporting still makes sense in month twelve.
The most common failure: inconsistent medium naming across platforms. One team tags Meta as "cpc," another tags it "paid-social," a third just uses "facebook." GA4 now sees three separate mediums for what is, functionally, one channel. Your paid social performance gets sliced into fragments instead of rolling up into one clean view.
Second failure, and this one's sneaky: forgetting to tag Klaviyo flows. Abandoned cart, welcome series, post-purchase, whatever it is. Untagged email links show up in GA4 as "direct" traffic, because the email client strips referrer data. So your direct traffic bucket balloons, looks suspiciously healthy, and nobody can explain why. It's not new customers discovering you by typing your URL from memory. It's unattributed email revenue.
Third: manual UTM entry. Someone hand-types tags into a link before a launch, misspells "campaign," and now that one ad's traffic vanishes into an unrecognized parameter bucket. Use a UTM builder or a templated spreadsheet with dropdowns tied to your locked taxonomy. It's boring infrastructure. It's also the difference between reporting you trust and reporting you have to caveat every time you present it.
Different channels need slightly different tagging logic.
Paid social (Meta, TikTok): use utm_content to identify the specific creative variant, not just the campaign. If you're testing three video hooks against the same audience, you want to see revenue per hook, not one blended number for the whole ad set. Our Meta reporting solution is built around exactly this: pulling creative-level performance instead of stopping at the campaign level.
Paid search (Google Ads): Google's auto-tagging (gclid) is generally reliable and easy to lean on, but keep UTMs running in parallel as a backup. If auto-tagging breaks or gets misconfigured on a landing page, you don't want your entire paid search channel to vanish from reporting until someone notices.
Email (Klaviyo, Mailchimp): don't just tag everything utm_campaign=email. That tells you nothing. Tag the specific flow: bfcm_flow_abandoned_cart, welcome_series_email_3, winback_60day. You want to know which flow drives revenue, not just that "email" did something.
Affiliate and influencer links: give each partner their own utm_source. Not "affiliate" as a blanket source for twenty different people. If you lump them together, you can't tell which partner is actually worth the commission and which one is dead weight. This one gets skipped constantly, usually because someone's managing affiliate links manually in a spreadsheet instead of through a system.
Tagging the link is step one. Getting from "GA4 shows a session" to "here's the actual revenue and CAC for that campaign" is a different problem, and it's where most brands hit a wall.
GA4's UI is fine for a single-channel view. It starts to break down once you're running paid social, paid search, email, and affiliate at the same time and need one blended picture. You end up exporting to spreadsheets, cross-referencing ad spend from four different platforms by hand, and hoping nobody fat-fingers a formula.
A warehouse-backed dashboard skips that step. Trivas runs on Amazon Redshift, joining UTM-tagged session data from GA4 with Shopify order data and ad spend across platforms. That gives you blended CAC and ROAS per campaign, not just per platform, which is the number that actually matters when you're deciding where next month's budget goes. Our BI reporting product is built specifically for this join, and our data dictionary lays out exactly how each metric gets calculated, so nobody's guessing at definitions.
This is also where clean UTM discipline pays for itself. Sloppy tags upstream mean someone has to manually clean and reconcile campaign names downstream, every single reporting cycle. Get the naming convention right once, and the dashboard just works.
Before your next link goes live, run through four steps:
Four steps, five minutes, and you avoid weeks of reporting confusion later.
If you're already tagging everything correctly and still spending hours reconciling spreadsheets across channels every week, that's a tooling problem, not a UTM problem. See how Trivas's dashboards handle it instead of doing it by hand every Monday morning.
explore more insights
3 min read
3 min read
3 min read
.d53b12e5.png&w=3840&q=75 "Trivas.ai")