Engagement rate means three things, so decide which one you mean

Engagement rate has no agreed formula, and that's the catch

Open rate and click rate have agreed sums. Engagement rate doesn't. One report blends opens and clicks, another adds replies, a social tool divides likes by followers, and Google Analytics uses it as the mirror image of its bounce rate. One name, several sums.

Why opens make a weak engagement signal now

Apple Mail loads images on the reader's behalf, so an open no longer proves anybody looked. Any engagement rate built mostly on opens is partly counting machines. Even Klaviyo's own engagement report for lists and segments sorts people by how often they open, so read it with that in mind.

So weight the signals a machine can't fake. A click, a visit to the site, an order, a reply. Keep opens as a tie-breaker, never the main test.

Define an engaged subscriber once, then use it everywhere

The useful version isn't a rate on a dashboard. It's a rule about a person: clicked, visited or bought inside a window you've chosen. Size that window to how often people buy from you. A coffee subscription and a sofa shop shouldn't share one.

• Campaigns go to the engaged segment by default, not to everyone
• People drifting out of it get a different email, not the same one louder
• People outside it for a long stretch go into a sunset flow and stop getting campaigns
• Every report uses the same definition, so the numbers line up

Mailbox providers watch how people treat your mail, so the rule protects the inbox as well as the report. It's the first thing we write down in a lifecycle strategy, and list segmentation is where it turns into the groups you actually send to.