The first app-install banner I ever set up was a disaster. Top of every page, full width, "show immediately," no dismiss cooldown. Within a week mobile bounce rate was up about four points and I couldn't figure out why until I actually opened the store on my own phone. The thing was shoving the "Discover What Inspires" hero halfway off the screen before anyone had read a word.
So let me save you that week.
App banners aren't a bad idea. They're a tactic that's almost entirely make-or-break on the configuration, and most of the advice online skips straight past that part.
Who this is even for
Be honest with yourself first. If you don't have a mobile app, a banner is pointless — a "smart app banner" only makes sense when there's an app worth installing. And even then, the people worth nudging are usually your returning customers, not someone who landed from a TikTok ad thirty seconds ago and has no idea who you are.
The math that makes it worthwhile: app users tend to come back more, and you get a push-notification channel you don't have to rent from Meta. That's the payoff. The cost is real estate and attention on a screen that's already cramped.
The mistakes that make people bounce
I've watched enough session recordings (Microsoft Clarity is free, use it) to see the same patterns over and over:
- Firing on the first pageview, for everyone. A brand-new visitor doesn't want your app. They want to see if your product is any good.
- Covering the content. A top banner that pushes the hero down, or a bottom bar that sits on top of the add-to-cart button, is worse than no banner.
- No memory. Shopper dismisses it, changes page, and there it is again. Nothing says "we don't respect you" quite like a popup that won't stay closed.
- Showing it on desktop. There's no app to install from a laptop. This one still surprises me how often it happens.
- Sending iOS users to Google Play. If your banner doesn't detect the operating system, half your links go nowhere useful.
What actually works
Here's the setup I've landed on after a fair amount of trial and error:
Bottom placement, and let it breathe. A slim bar pinned to the bottom stays out of the way of the hero and product imagery. If your theme has a sticky add-to-cart, mind the overlap — a little padding around the banner (the "floating" look) usually reads cleaner than edge-to-edge.
Delay it, or trigger on scroll. Don't show it until someone has scrolled a bit or spent a few seconds. That tiny wait filters out the immediate bouncers and only bothers people who are actually looking around.
Remember the dismissal. When someone closes it, keep it closed for a week or so. A dismiss cooldown is the single setting that separates "helpful" from "nagging."
Detect the OS. iPhone users get the App Store link, Android users get Play. This should be automatic, not something you manually pick.
Target where it shows. You probably don't want it on the cart or checkout, where it competes with the one action that matters. Homepage and collection pages are fine; checkout is not.
If you'd rather not wire all of that up by hand, this is more or less exactly what we built Scurix App Download Banner to do — OS detection, placement and behavior controls, a dismiss cooldown, all no-code. But the principles above hold whatever tool you use.
A configuration I'd actually ship
If you want a starting point that won't get you angry emails:
- Position: bottom, padded (not full-bleed)
- Show delay: a few seconds, or on first scroll
- Dismiss cooldown: 7 days
- Devices: mobile only
- Pages: everywhere except cart and checkout
- OS detection: on
Then leave it alone for two weeks before you touch anything.
Actually measure it
This is the step everyone skips. Before you decide the banner "works," send an event to GA4 when it's shown, clicked, and dismissed. What you want to know is the click-through rate on the banner and — more importantly — whether the pages showing it kept the same bounce and conversion rates they had before.
If bounce climbs, you were too aggressive. Pull back the timing or the placement. If nobody clicks, the offer isn't compelling (or your audience is mostly one-time buyers who'll never install an app, which is also useful to learn).
A banner is a small thing. But it sits on top of every mobile session, so a small mistake gets multiplied by every visitor you have. Treat it like the high-traffic surface it is, ship the conservative version first, and let the data tell you how far you can push.
