Q2 CVR Collapse: Compounding issues drove the NC decline: three are now confirmed with data and we have fixed or are mitigating.
1. Media auction inflation (confirmed) Meta CPM more than doubled at peak (Apr 27: $47.82 vs. March avg: $23.29), compressing impressions by ~65% at flat spend. CPC rose ~25%. CTR actually improved, which partially offset the CPM increase, meaning the creative wasn't the problem (although the current creative is very MOF/BOF instead of TOF, which we are working to fix). This is auction-level cost pressure, and it is partially easing (June CPM is back to $31.91). A creative refresh with the new strategy and audience work are the only levers to continue to move these CPMs down. More on that below.
2. Rise-2 Variant LP Testing Increases CVR Degradation (TBD - in the process of confirming)
Reminder: Key MUD Funnel for OG Subscription: LP (”Rise-2” for the OG) > OG PDP > Cart > Checkout
The Rise-2 landing page is Mud’s main LP and has been since 2025. In 2026, with the return of the Sr. UX Designer from Pat leave in March and the increasing pressure to hit NC goals, we restarted the testing roadmap on Rise-2, laying out both “big swing tests” and incremental tests to improve CVR. Testing is done through intelligems redirects on the Replo PDP (cloned versions or completely new Replo LPs). We began testing 1x per week (ex: Listicle LP vs. Rise-2) and then ramped up testing as we increased our ad diversity (ex: Rise-2 vs. Influencer specific Rise-2 variant for the influencer’s traffic, Rise-2 with CTA copy test, Rise-2 with hero image test). We began stacking tests (1-2 per week with specific traffic carved out to get its own tests. With the increasing use of intelligems, we began to see erosion of a key metric in the Mud funnel that is not visible from the shopify funnel CVR graphs.
We began to see a deterioration of LP > PDP (this CVR is incredibly important because the user needs to hit the PDP to get the email popup + more information about the actual product, which will lead them to convert later on if not in that session). This deterioration is specific to Rise-2 variants accessed through intelligems redirect tests.
An Important Point: users who did click from the Rise-2 LP to the PDP had the same CVR to ATC and checkout that the normal Rise-2 without intelligems redirect had.
So - traffic going to Rise-2 had a degradation in this conversion rate if there was an intelligems redirect running on the page. When redirect tests were slowed down/ turned off. CVR would stabilize.
With 10% / 90% tests - In intelligems, we saw slightly negative results that would have trouble getting to stat sig. So we increased the proportion of the traffic in the intelligems test to 50% in order to get reads faster and test at a higher velocity. This exacerbated the problem - pushing all traffic into the redirect.
We started to see CVR tank, which is what led us to try to test more, faster. It was a negative feedback loop, with LP > PDP CVR getting worse each week despite strategically designed tests that should have increased the CTR to PDP from the LP.
Diagnostics: pass- $100m
https://growth-diagnostic.pages.dev/funnel
https://cac-cvr-diagnostic.pages.dev/
6/11 Update: Kill-switch the Intelligems code in the Shopify theme immediately. See if CVR improves. If test routing is the cause, LP→PDP CVR and ECR recovers toward April levels within a day. This is the fastest way to separate what's fixable immediately from what requires the creative/auction rebuild.
3. Email Popup Disabled 4/22-6/4 (confirmed) With the launch of the Straw Matcha LTO, there was an ask from the internal team to turn off the popup from PDPs. This decision was made without getting sign off from the growth team due to the silos within the teams (brand operates independently to manage the email agency, consulting/informing the growth team on email and popup for promotions/flows). Why was this not discovered sooner? The Alia popup (recently implemented) rapidly increased the capture rate of email+SMS so entries into the email program being elevated hid the issue that the PDP was no longer a Popup feeder.
6/4 Update: Issue discovered and corrected. Silo killed, all decisions are informed by growth team to make sure there are no funnel / revenue repercussions. Popup signups by source tracked on a weekly basis to measure for changes.
6/1 Update: Issue discovered and corrected. Silo killed, all decisions are informed by growth team to make sure there are no funnel / revenue repercussions. Program is now optimized for NC scale not solely ROAS.