Web Push Notifications for Ecommerce

Web push notifications for ecommerce: a practical guide

Web push is the most underused channel in ecommerce retention, and the reason is almost always the same: brands enable it badly once, watch everyone block them, and quietly conclude it doesn't work.

It does work, but it's unforgiving in a specific way. Get the opt-in wrong and you don't just lose the subscriber, you lose the ability to ever ask again. A blocked browser permission is permanent. That single fact should shape everything about how you run the channel.

Why it's worth the trouble

Three things make push structurally different from your other channels.

First, it costs effectively nothing per message. No per-send fee like SMS, no deliverability score deciding whether it lands. Second, it bypasses the inbox entirely, so it isn't competing with the fifty other emails your customer received today. Third, and most importantly, it reaches people who will never give you an email address or a phone number. A shopper who won't fill in a form will often click "Allow" without breaking stride. That audience is otherwise completely invisible to you.

For a brand with meaningful traffic and a modest list, push can quietly become the cheapest incremental revenue available.

The opt-in is everything

Here's how most brands destroy the channel. They switch on the native browser prompt, it fires the instant a first-time visitor lands on the homepage, the visitor reflexively hits "Block," and that's the end of the relationship. Permanently.

The fix is a two-step approach. Show your own branded primer first — a soft, on-brand panel that explains what the shopper actually gets: restock alerts, early access to drops, price-drop notifications. Only when they say yes to your panel do you trigger the real browser prompt. If they decline the primer, nothing happens, no permission is burned, and you can ask again later.

Timing matters just as much. Asking three seconds into a first visit is asking for commitment before you've given a reason. Trigger the primer on demonstrated intent: two or three product views, real time on page, an add to cart.

What push is genuinely good at

Push is short, interruptive and has no room for storytelling. Use it where immediacy is the value:

• Back-in-stock alerts. The subscriber explicitly asked to be told. This is about as close to guaranteed intent as ecommerce gets.

• Price-drop alerts on a product they actually viewed.

• Cart and browse abandonment, as a free recovery layer running underneath email.

• Drops and launches, where speed decides who gets the product.

• Final-hours urgency on a closing sale.

What it's bad at is long persuasion and daily promotional noise. Push has no equivalent of the polite email unsubscribe. People just block you.

Segment it, or train people to block you

The lazy version of push is one broadcast to everyone who ever clicked "Allow." The browser gives you behavioural data, so use it. Target by the product someone viewed, what's sitting in their cart, the category they keep browsing, how recently they visited. A restock alert sent only to the people who looked at that exact product reads as a service. The identical message sent to your whole subscriber base reads as spam.

Frequency discipline is not optional

This bears repeating because it's where brands fail. Every push you send carries a small risk of a permanent block. That changes the maths entirely compared to email, where an over-send costs you a bit of engagement. Fewer, sharper, better-targeted notifications will always out-earn a high-volume program, because the audience survives to be marketed to next month.

Run it as part of the system, not beside it

Push earns its keep as the safety net in a coordinated lifecycle. A shopper abandons a cart. The email lands and goes unopened. They're not on your SMS list. A push the next morning is the only remaining way to recover that sale, and it costs nothing to send. That's the logic: push isn't another megaphone, it's the channel that catches what the others miss.

Which means it needs to share the same customer data and the same frequency governance as everything else. A customer shouldn't get an email, a text and a push about the same sale inside an hour. The rules that prevent that are the same ones we set out in email and SMS orchestration.

What to measure

Track opt-in rate (the primer is the lever), click-through rate, revenue attributed to push, and — as your early-warning system — your block and opt-out rate. If blocks are climbing, you're sending too much or targeting too loosely, and you're burning an audience you cannot get back.

One more that's easily missed: measure push-assisted revenue, not only push-attributed revenue. Because push so often acts as the second or third touch in a recovery sequence, last-click attribution will systematically undervalue it. The better question is whether your overall cart recovery rate improved after push went live, not just how much revenue push claimed for itself.

The economics, plainly

It's worth being concrete about why push is worth the setup effort. Email costs almost nothing per send but competes for attention in a crowded inbox and depends on a deliverability reputation you have to actively protect. SMS gets read almost immediately but carries a real per-message cost and the strictest consent rules of any channel.

Push sits in a genuinely useful gap between them. It has SMS-like immediacy with email-like economics, and it requires no personal data from the customer at all — just a browser permission. For a brand with decent traffic but a modest list, that combination frequently makes push the cheapest incremental revenue in the entire stack. It is not a replacement for either channel. It's the layer that catches the people and the moments the other two structurally can't.

Setup: what good actually looks like

The technical requirements are modest — an HTTPS site, a service worker, and a push platform — which is exactly why push so often ends up as a plugin somebody switched on and nobody owns. That's the failure mode to avoid.

Treat it like a real channel. Someone owns the opt-in rate. Someone owns the flows. The segments come from the same customer data as your email and SMS, not from an isolated tool with its own idea of who the customer is. And the frequency rules are shared, so push is governed by the same calendar as everything else rather than firing whenever a separate dashboard suggests it. Run it that way and push stops being a plugin nobody owns, and starts being an asset that quietly compounds.

Frequently asked questions

Do web push notifications actually work for ecommerce?

Yes, particularly for back-in-stock alerts, price drops and cart recovery. They cost almost nothing per send and reach shoppers who never provide an email address or phone number. The results depend almost entirely on the opt-in approach and on frequency discipline.

How do you increase push notification opt-in rates?

Never fire the native browser prompt at a cold visitor. Show a branded two-step primer explaining the benefit, triggered after the shopper has shown intent, and only request the browser permission once they've accepted your primer.

See how we build push into a coordinated retention program with our push notification services.