I want to tell you about the last time a notification was right.
Mine was a Tuesday in February. It was raining. I was walking into a Starbucks. My phone buzzed once, lock screen lit up, and the banner said "$14, you're here." A Starbucks card I'd loaded two months earlier and forgotten about. I swiped up, the barcode came up, the barista scanned, and I had a coffee.
The notification, a single line of text on a rainy Tuesday, was correct.
I want you to try to remember the last time one of yours was. Scroll back a day and count. I'm genuinely asking.
The number isn't zero, usually. But the number is much smaller than the number of notifications you got. Most notifications are a tax you pay for the one that was right. The app economy has converged on a quiet default: ship the alert, let the user mute us later if it becomes a problem. Which is a reasonable strategy from the app's perspective and a slow catastrophe from yours.
The app I'm building is at war with that default. Not gently, structurally. The whole reminder architecture is built around a single strange idea: the app should try not to notify you. Its first move in every single case is to ask whether now is the wrong time, not whether now is a possible time.
Here are some of the reasons, at any given minute of your day, that CardCue Pro may decide not to tell you about a card you own:
- You're already inside the store. (It fires approaching, not arrived.)
- You're driving past the store on a highway at 50 mph, not stopping.
- You're at home, and home is a quiet zone you've configured.
- Your phone has Do Not Disturb on, or any Focus mode on.
- It's raining hard outside and the card is for a car wash.
- You were at this same store yesterday and the day before and you're not particularly loyal to it, you pass it every day.
- You said "not now" to this card three notifications ago; the app remembered and backed off for a week.
- There's another reminder about to go out in the same five-minute window and the app decided this one matters less.
- It's Sunday morning and you probably haven't had coffee yet.
Any one of those is enough to cancel the alert. Not "delay" it. Not "schedule it for later." Cancel it. The moment passed; the card waits in the drawer for the next one.
If you're doing the math, you've already noticed something: most of these are not product features, they're absences. They're alerts that didn't fire. Which is a hard thing to demo, sell, or put on an App Store screenshot. The best moments of CardCue Pro are the ones that would have been bad moments on a dumber app, and then weren't anything at all.
People who try the app and don't know what it's doing sometimes ask if the notifications are working, because they haven't seen any. That is, genuinely, one of the best reactions I get.
Let me say something philosophical for a second, because this is the part that matters.
Every notification is a promise you make to a human. The promise is: I have broken into your attention for a reason, and you will not regret it. If you keep that promise, you get to break in again. If you break it, you get muted, uninstalled, and, more quietly but more importantly, you make the person a little more cynical about the next notification they get from anyone else.
Most apps don't think of it this way. Most apps have one setting for alerts: on or off. They assume that because they have permission to notify, they should use that permission proportional to the number of things they have to say. "We have seven updates for you, so let's send seven notifications." That is the mind of an app that does not believe it has to earn the next one.
Cue is trying to believe it. Hard. The way it tries is five stacked gates (location, history, conditions, personalization, throttle) and roughly twelve input signals, from your calendar and from WeatherKit and from your own behavior in the app. The technical write-up of how that actually works, for the people who want it, is the Geolocation Brain press kit, five gates, twelve signals, and the quiet math that decides whether a minute is The Right Minute.
But that's not what I actually want you to take away. What I want you to take away is: the right notification at the right moment is worth a hundred wrong ones. That's not a poetic claim. That's the real exchange rate. One great banner at a Starbucks on a rainy Tuesday is worth a long run of silence, and if you're willing to do the silence work, the banner eventually lands.
Most of the apps on your phone are not willing to do the silence. That's why you're tired.
The Starbucks card on the rainy Tuesday is not the app's most important feature. The app's most important feature is everything it didn't say on all the other Tuesdays.