Open rates are a poor compass for lifecycle work

Open rate survived because it is easy to put in a slide. Vendors compute it. Stakeholders recognise it. It moves when you change a subject line. None of that makes it a compass for lifecycle messaging analytics.

What an “open” usually is

On email, an open is often a pixel that loaded, or a privacy proxy that loaded it for you. On push, an open is frequently a tap on a lock screen that never became a session. On in-app, teams sometimes count any render inside a bottom sheet as attention. These events are not comparable, yet they get averaged into one percentage that then governs send volume.

In Signal Path Studio we force a split: delivered, rendered, glanced, acted. If your warehouse cannot support that split yet, say so in the report. Pretending the vendor column is “attention” is how quiet-hour policies get ignored.

When a rising open rate hides a dying journey

We keep a sanitised example from a Bangkok consumer app. After a copy change, push open rate rose. Unsubscribes also rose. Seven-day return among previously active users fell. The copy was more alarming; people tapped to dismiss anxiety, not to continue a task. The campaign recap still led with the open bump.

Journey Ledger would have shown interference with an in-app card that already explained the same task. The tap was not incremental. It was a second reminder of something the product already said.

What to put on the wall

We ask Atelier students to pin four numbers next to any journey: delivery success among eligible users, render among delivered, a downstream action that the product already cares about, and the same action in a holdout. Open rate can live in an appendix. It should not set the agenda.

If legal will not allow a holdout yet, say that the compass is incomplete. A missing counterfactual is more honest than a confident open percentage.

Back to field notes