We stopped relying on install-level attribution and started preserving UTM tags through the web funnel. That change made it much easier to trace churn to the campaign, creative, or placement that brought the user in.
On mobile you often lose creative-level detail after ATT or when a user moves through an app install flow. On web I could keep the full UTM chain attached to the customer record and then run downstream cohort analysis in Mixpanel. That let us see which creatives produced quick churning users versus the ones that produced long-term subscribers.
The lesson: preserved UTMs don’t fix a bad product, but they show which ads are creating toxic traffic so you can pause or rework them quickly.
What data points do you keep alongside UTMs to make churn attribution more reliable?
I persist the UTM on the user record and surface it in our subscription table.
When something looked off I pulled the JSON from our web generator and traced the UTM through the onboarding flow.
That simple mapping saved a lot of guesswork when assessing campaign quality.
Preserve utm and tie it to every downstream event.
I store campaign, creative, and experiment ID on the user object so I can slice churn by those dimensions.
Using a web-first funnel lets you keep that info intact and see which ad variants produce short lived subscribers.
I keep utm, landing page variant, and first plan chosen.
When churn spikes I check which landing pages and creatives they came from.
save utm on signup then use it to slice churn
UTMs plus these fields: creative id, landing page id, experiment id, and first purchase price.
That combo lets you separate bad creative from bad pricing. On web you can preserve that chain end to end and run quality checks like retention curves per creative. If a campaign brings many installs but poor LTV you either change the creative or pause it.
We attach utm, ad id, and the landing copy version to the user profile.
Then I run a monthly report that flags campaigns with high churn within 30 days.
We store utm and creative id and it helped us find one TikTok ad that drove lots of refunds.
Also link payment method and initial price. Those often explain churn more than channel alone.