What Counts as an Engaged Email Segment in 2026
An engaged segment is a rolling definition of who has shown they still want your email, and almost every deliverability problem in ecommerce traces back to a brand that never wrote one down. Here is the version that works, and the rule for which campaigns are allowed to reach which window.
Use Click, Not Open
Apple Mail Privacy Protection pre-loads images, which means a meaningful share of your recorded opens belong to a server rather than a person. A segment built on opens will quietly fill with profiles that have not looked at your email in a year.
Build the segment on clicks, and use opens only as a secondary signal inside a window you have already narrowed.
Add two behavioural signals that outrank both: placed an order and viewed a product. A customer who bought sixty days ago without clicking an email is engaged with your brand even if they are not engaged with your inbox, and excluding them is how brands accidentally suppress their best buyers.
The Three Windows
Thirty days. Clicked, viewed a product, or ordered in the last month. This is your safest audience—the one you use to warm a domain, test a new send slot, and retreat to if complaint rates rise.
Ninety days. The working segment for ordinary campaigns. On most DTC lists, this is between a third and a half of total profiles, and it produces the large majority of campaign revenue.
One hundred and eighty days. The widest audience any promotional campaign should reach, reserved for genuine launches and the two or three biggest sends of the year. Beyond one hundred and eighty days, you are not marketing; you are testing whether a mailbox provider is paying attention.
Past twelve months without a click, the profile belongs in a sunset flow, not in a campaign.
Why the Windows Are the Deliverability Control
Google’s guidelines require bulk senders to keep the Postmaster Tools spam rate below 0.10% and never to reach 0.30% .
Because that is a rate, the arithmetic is entirely about audience composition. A complaint count that is harmless on a send to 20,000 engaged profiles becomes a threshold breach on a send to 100,000 where the extra 80,000 have not clicked in a year, because those 80,000 complain at a much higher rate.
Widening the window is the single most common cause of a sudden placement drop, and narrowing it is the fastest fix.
Validity’s benchmark showed inbox placement falling from just under 87% in February 2024 to 82.3% during the Q4 peak season, while spam placement almost doubled. Most of that swing is brands widening their audience for the season.
One More Segment Worth Building
Separate never-purchased from lapsed-purchaser profiles inside every window.
They want opposite messages. The first needs a reason to try you; the second needs a reason to come back. Sending both the same campaign wastes the send on one of them.
Attentive found that 81% of subscribers who opted out during a peak week had never made a purchase, and another 15% had bought once. That is a direct measurement of what happens when the never-purchased group gets treated like customers.
One practical note on maintenance: these are rolling definitions, which means a profile moves between windows on its own as time passes, and that is the point.
What breaks them is a brand that builds the segments once and then starts sending to a saved static list because it was easier on the day.
Audit campaigns quarterly and check that each one went to a dynamic segment rather than a snapshot. Static lists are how a clean programme slowly turns into a list you cannot trust.
Build the windows in October, freeze the definitions before peak, and never edit one mid-campaign.
The wider segmentation work, including purchase-behaviour and lifecycle splits, is covered in ecommerce email segmentation, and we build and maintain these segments inside client accounts as part of our Klaviyo-first email marketing service.
