For five years, mobile marketing rebuilt who the user is. Almost nobody rebuilt when they are reachable. The when never broke, because it never needed an identity.

For five years, mobile marketing has been one long effort to rebuild what ATT took away. SKAN, clean rooms, first-party data, probabilistic attribution. All of it trying to answer one question: who is this user.
A different signal sat untouched the whole time. Not who the user is, but what they are doing right now. It never needed an identity, so privacy never took it away.
The scramble after ATT was about identity. The IDFA went dark, attribution got fuzzy, and nearly every dollar of effort went into recovering the ability to know who a user is and where they came from. Clean rooms, first-party data strategies, sandbox proposals, probabilistic matching. All of it identity recovery.
It made sense. Identity was the signal that broke, so identity was the signal everyone rushed to fix.
But identity is also the signal that keeps eroding. Every recovery method sits one policy change away from getting harder again. The ground under "who" has not stopped moving.
There is a whole class of signal that the privacy changes never touched, because it was never built on identity in the first place.
Not who the user is. What they are doing right now, in the real world. Is the phone flat on a table. Is the person walking. Is this a focused moment or a distracted one.
These signals come from the device itself: motion, orientation, screen state and others. They are inferred on the phone, never tied to an identity and never sent anywhere. There was nothing for ATT to take, because none of it depended on knowing who anyone was.
Context is privacy-durable by construction, not as a compliance feature bolted on afterward. As identity signal keeps degrading, context signal does not, because it never leaned on identity to begin with.
Here is the split most teams have not drawn.
Identity answers who to acquire and who to talk to. Context answers when the person you already have is actually reachable.
The industry spent five years rebuilding who. Almost nobody was rebuilding when. And the when got easier to act on over the same stretch that the who got harder.
This is not a replacement argument. You still need identity to run acquisition. Context does not target better, it times better. The two answer different questions, and one of them had a clear runway the whole time everyone was looking the other way.
Timing is not a soft lever. It shows up directly in the numbers.
About 53% of push notifications land on a phone lying flat and unattended, where nobody sees them. When the phone is upright and in a hand, open rates run 2 to 3 times higher. Same message, different moment.
That gap is retention and revenue that identity work can never reach, because it was never a who problem in the first place.
This is why context reads as infrastructure, not a feature. Web analytics, mobile attribution and the cloud warehouse each became a layer an entire category was built on. Each also got reshaped by the same privacy shifts that hit identity.
On-device context is the next layer, and it happens to be one of the few that those shifts leave intact. The raw signals are noisy on their own. On-device machine learning turns them into an inference about the user's real-world context. That inference is probabilistic. It informs a decision, it does not determine an outcome. And it stays on the phone.
ContextDecision and ContextPush are what you build on that layer today. ContextDecision times the in-app moment, when to show the offer or the next step. ContextPush times the notification, so it lands when someone is actually looking. Feeding context further upstream into acquisition itself is an emerging direction rather than a settled one.
Identity signal will keep being contested, and teams will keep spending to recover it. That work is not wrong. It is just the crowded half.
The other half sat open the whole time. The when did not break, because it never needed anyone's identity. It is the rare lever that got clearer while everything around it got harder.
Signal loss broke targeting. It did not touch timing.