Progressive
web app development
in Santa Monica

Some tasks fail the moment a screen dims and locks. Timing is everything. A checkout screen showing a code for someone to scan, a recipe open on a counter, a workout timer running mid-set: all of them break if the device sleeps halfway through. A progressive web app can ask the browser to hold the screen awake for exactly as long as that task needs, and no longer.

The request is scoped to the page, not the whole device. Switching to another tab, minimizing the browser, or locking the device manually all end the hold immediately. Coming back to the page does not automatically restore it. Nothing happens by default. The app has to notice that the page became visible again and ask a second time, or the screen will happily dim on the very next attempt.

Support is not universal. It varies across browsers and versions, so a design relying on it needs a real fallback path. Detecting whether the capability exists before relying on it, and simply doing nothing when it does not, beats a broken feature that throws an error a person never sees but definitely experiences as a screen going dark at the wrong moment.

Holding the screen awake has an obvious cost. Battery drains faster the whole time the hold stays active. A task that runs for a few seconds is not worth thinking twice about. A task that could run for an hour, like a long workout session or a slideshow left running in a store window, deserves a visible indicator that the screen is being kept on deliberately, so nobody wonders later where the battery went.

Releasing the hold the moment the task actually finishes matters as much as requesting it in the first place. Timing cuts both ways. A payment screen that keeps the display awake long after the transaction completed is just draining a battery for no reason. Tying the request and the release to the same clear moments in the workflow, rather than leaving it running for the whole session, keeps the feature doing exactly the job it was added for.

None of this needs a permission prompt shown to the person using the app, which makes it easy to reach for without much thought. That ease is deceptive. It deserves a deliberate decision about where it actually helps, not a reflex used everywhere. It should not become a default switched on for every screen just because the code to do it is short.

A browser only shows the push notification permission prompt once per site, on most platforms. That single shot stays fixed until a person clears the decision in browser settings or reinstalls entirely. Timing matters more than wording. Getting the moment right is one of the most consequential decisions inside a progressive web app. The wording matters less.

Asking on the very first visit produces one of the lowest acceptance rates of any pattern. A person has not done anything inside the app yet. Nobody has decided the app is worth hearing from. A cold request rarely works. It comes from an unfamiliar product, and it tends to get an automatic decline. That decline is effectively permanent. Almost nobody goes back to change it later in browser settings.

A better pattern asks after a specific, relevant action. Completing a booking that needs a status update is one example. Finishing a form that will take time to process is another. So is reaching a point in a workflow where a notification would replace a person having to come back and check manually. The request feels like it follows logically from something the person just chose to do. It does not arrive out of nowhere.

A soft ask can come before the real one. This is an in-app message. It explains what notifications would be used for and offers a plain button to proceed. It lets a person opt out without burning the one real browser prompt. Anyone who declines can be asked again later without penalty, since the actual browser permission was never triggered. Anyone who accepts sees the real browser prompt already primed to say yes.

What the notifications will actually contain matters as much as when they are requested. Scope creep is tempting. Permission granted for order updates should not turn into weekly marketing messages. That kind of drift teaches a person to distrust every future prompt from any installed web app, not this one alone. Keeping the content narrow and predictable protects the channel for the messages that actually need it.

Once permission is granted, give people a visible way to turn notifications back off inside the app itself. No one should hunt through settings for it. That keeps the relationship a two-way choice rather than a one-time trap. Products that make opting out easy tend to see fewer people revoking permission at the browser level entirely. That level blocks every future message, wanted ones included.