How Menza traced a 48% drop in email click rates to a single timing setting buried in a Klaviyo flow migration

The Challenge

HeyLo! had just finished rebranding all their Klaviyo email flows. The new flows were live, sending 4x the volume, reaching 3x more recipients. On paper, the migration looked like a success.

But something was off. The team suspected engagement had dropped since the rebrand, but couldn’t pinpoint why. Comparing flow performance meant normalising for different time periods, isolating variables across dozens of email actions, and digging into timing structures that weren’t visible in standard Klaviyo reporting. The kind of analysis that would take hours—and might still miss the real issue.

How Menza Helped

They asked Menza to compare the old and new Browse Abandonment flows using matched date ranges: April 7–May 7 for the original, June 17–July 17 for the rebrand.

Menza pulled performance data across both periods, standardised the comparison, and surfaced the gap: the old flow hit a 10.03% click rate. The rebrand was stuck at 5.21%—nearly half the engagement, despite sending more emails.

Then Menza went deeper. It analysed the flow structures, message sequencing, and delay settings for each email action. The culprit: the old flow used 2-hour delays between emails. The rebrand had simplified everything to 1-hour delays. The 2-hour emails consistently outperformed, hitting 10%+ click rates. The 1-hour emails plateaued around 5%.

Menza didn’t stop at diagnosis. It built an A/B testing framework—50/50 split, 4-6 week duration, statistical requirements defined—so the team could validate the fix before rolling it out across all abandonment flows.

The Outcome

In one conversation, HeyLo! went from “I think engagement dropped” to knowing exactly why, exactly where, and exactly how to fix it. A timing decision that seemed trivial—1 hour vs. 2 hours—was responsible for cutting click rates in half. The A/B test framework was ready to deploy, with a projected 92% improvement in click-through performance if the hypothesis held.