Your Tuesday retention review has a new line item. The Monday blast that went to all 120,000 web push subscribers on your Shopify Plus store delivered cleanly in January. This week, a chunk of it never arrived, and the dashboard shows delivery sagging a little more each send. Meanwhile the cart-abandonment trigger you set up two years ago keeps doing what it always did: 4% of your send volume, a third of your push-attributed revenue.
That gap is not a reporting glitch. It is policy. Google spent 2026 rewiring how Chrome and Android treat notifications, and triggered push notifications are the winning side of every change: blasts lose reach, triggers gain it. This post walks through what changed, the deliverability and revenue math behind it, and the first trigger campaigns to ship if your program is still blast-first.
Google spent 2026 splitting your send list in two
Three platform changes, one direction.
Chrome now rate-limits pushy senders. Starting January 2026, Chrome began rolling out Push API message rate limits for sites that send heavily relative to the engagement they earn. Chrome scores every sending origin daily on three factors: push messages sent per time subscribers spend on your site, permission prompts shown per time on site, and user engagement with your site itself. Fail the test and Chrome web push rate limits cap your sends, with requests over the cap bounced with an HTTP 429 error. Google’s own framing is blunt: senders of “timely, relevant, and engaging notifications” will not notice. Megaphones will.
Android 16 silences promotional push. The Notification Organizer uses on-device AI to read notification content and auto-silence what it classifies as Promotions or News, both filtered by default. Classification happens per notification, not per app, so a promo blast from your app gets buried even if your transactional messages sail through.
Android 17 gives live notifications premium placement. Announced at I/O 2026 and now rolling out, Live Updates pin genuinely live notifications to the top of the notification drawer, the lock screen, and a status-bar chip. Google’s developer documentation is explicit about what does not qualify: “ads, promotions, chat messages, alerts.” The same company that is silencing your promo blast built a premium surface for your order-tracking update.
RCS Business Messaging picked up the same posture at I/O 2026, with verified-sender branding and daily caps for low-reputation promotional senders. The pattern holds everywhere: the platforms have stopped treating all notifications as equal, and they have chosen the triggered side.
What separates a triggered push notification from a blast with better timing
A working definition, because the industry blurs these constantly. A triggered push notification fires for one subscriber because that subscriber did something. It has an entry condition (the event), per-subscriber state (their cart, their browsed category, their last order), and an exit condition (they purchased, they clicked, the window expired). Behavior-based push notifications are campaigns built from these per-subscriber events rather than from a calendar.
That definition excludes more than you might think:
| Campaign class | What fires it | What stops it | How platforms treat it in 2026 |
|---|---|---|---|
| Scheduled blast | A calendar date | Nothing; it sends to everyone | Throttled when volume outruns engagement; silenced as Promotions on Android |
| Segmented blast | A calendar date, sent to a filtered list | Nothing per-subscriber | Better engagement math, same structural risk at volume |
| Triggered campaign | A subscriber’s action (cart abandon, price drop on a browsed item) | An exit condition: purchase, click, or expiry | Low volume per subscriber, engagement-rich; passes Chrome’s test by design |
| Live update | An ongoing, user-initiated event (order out for delivery) | The event ending | Premium placement on Android 17’s lock screen and status bar |
A daily “deals” send with smarter timing is still a blast. Scheduling software cannot change its class; only an entry condition tied to subscriber behavior can. Triggered notifications in PushEngage are defined exactly this way, as event-fired flows with entry and exit conditions rather than send-time optimizations.
The deliverability math: how Chrome decides you’re a megaphone
Chrome’s rate-limit scoring is a fraction you can manage. The numerator is how much you send. The denominator is how much attention your site earns: site engagement score plus the minutes subscribers actually spend there, calculated daily per sending origin.
Blasts inflate the numerator by definition. Every subscriber gets the send, including the long tail who have not visited in months and hold your engagement denominator near zero. Triggered campaigns invert the ratio: the only subscribers who get messages are ones who just visited, browsed, or carted, meaning their engagement is fresh by construction.
The penalty ladder makes the downside concrete. One flagged day earns a 1-day rate limit. A second flagged day extends it to 7 days. A third makes it 14, and the slate only wipes clean after 42 consecutive well-behaved days. A capped site is limited to 1,000 push messages per minute, and everything past the cap returns HTTP 429.
Run the scenario on a 100,000-subscriber store. A daily blast program that trips the ladder in mid-November spends Black Friday capped at 1,000 messages per minute, watching its flash-sale push trickle out over hours instead of seconds. Illustrative math: if push recovers $40,000 of cart revenue in a normal sale week, a throttled week that halves delivery puts roughly $20,000 of recovered revenue at risk. The trigger-first program next door never appears on the radar, because Chrome web push rate limits target the megaphone pattern, not the volume of genuinely engaged senders.
Two honest caveats. The limits apply to the background Push API, not the Notifications API, and Google expects nearly all sites to be unaffected. But “nearly all sites” is measured across the entire web. High-volume eCommerce senders running daily blasts are precisely the cohort the policy describes. PushEngage has sent over 100 billion notifications across 150+ countries, and the senders most exposed are the ones whose entire program is a calendar.
Revenue per send: the economics of triggered vs. blast
Engagement multipliers are everywhere in this category; revenue math is rarer. PushEngage’s own platform data puts open rates on personalized triggered campaigns at roughly 4× untargeted sends. The number that should reach your budget review, though, is revenue per send.
Illustrative, on 1,000 sends each:
| Program | Gönderimler | CTR | Conversion on clicks | Orders | Revenue at $80 AOV |
|---|---|---|---|---|---|
| Daily blast | 1,000 | 1.5% | 5% | <1 | ~$60 |
| Cart-abandonment trigger | 1,000 | 12% | 25% | 30 | ~$2,400 |
The spread is not subtle, and real-world cases land in the same direction. Charlton Athletic, the English football club, runs cart-abandonment triggers through PushEngage at a 7.17% conversion rate per send. Saleduck’s triggered drip campaigns earn 400% higher open rates than its broadcast sends. Because PushEngage reports revenue at the notification level, both teams can put a dollar figure next to each campaign rather than defending clicks. That is the difference between owning a line item and explaining one.
The 2026 asymmetry sharpens this further. A blast’s downside used to be unsubscribes. Now it also spends deliverability: every low-engagement send nudges your Chrome score toward the ladder and trains Android’s classifier to file you under Promotions. The blast does not just earn less per send; it taxes every future send. Trigger campaigns carry no such tax, and after the throttle threat the safest high-volume strategy left is making your triggered layer bigger.
Which events deserve trigger campaigns (and which don’t)
Not every event earns a campaign. The events worth automating pass three tests:
- Intent signal. The action predicts a purchase or a churn. A carted product is intent; a scrolled homepage is noise.
- Perishable window. The message is worth more now than tomorrow. Cart contents go stale; price drops get bought by someone else.
- Per-subscriber state. The message can name the thing: the product, the price, the category. If every recipient could receive the same text, it is a blast in disguise.
For an eCommerce store, six events pass all three: cart abandonment, browse abandonment, price drop, back-in-stock, welcome (the opt-in is the event), and win-back (the absence of an event, crossing a threshold). These six are also where behavior-based push notifications stop being an engineering project: each ships as a ready template in PushEngage Workflows, so your team configures timing and copy instead of architecting event pipelines. Browse abandonment deserves particular attention because it covers shoppers who never carted at all; our browse abandonment campaign strategies post covers how to define the audience without over-firing.
If you want the full menu beyond the core six, including saved-item alerts, cross-sell, and review requests, the eight trigger campaign ideas for eCommerce post walks each one with examples.
Exit criteria, frequency caps, and quiet hours now protect your reach
Campaign hygiene used to be courtesy. Under engagement-scored deliverability it is defense, and each control maps to a specific platform mechanism.
- Exit conditions stop a sequence the moment the subscriber purchases, clicks through, or the window expires. Every message a finished sequence keeps sending is pure numerator in Chrome’s math: volume with no chance of engagement.
- Frequency caps put a per-subscriber ceiling on total sends per week across all campaigns. Caps are what keep three overlapping triggers from re-creating blast volume one subscriber at a time.
- Quiet hours hold messages that would land at 3 a.m., when they get dismissed unread. Dismissed-on-arrival is exactly the low-value pattern Android notification silencing learns from.
Useful defaults: every triggered campaign ships with an explicit exit condition, non-triggered sends capped per subscriber per week, and quiet hours matched to each subscriber’s timezone. Sloppiness here shows up first as unsubscribes; our guide on how to reduce your push notification unsubscribe rate covers the warning signs before the platforms make the decision for you.
Triggered campaigns are only as good as the segments they fire on
A trigger answers when a message fires. A segment answers who is eligible. Skip the second question and a price-drop trigger fired list-wide is just a blast with a fancier entry condition.
Run the comparison. A price-drop notification sent to everyone announces a discount to 100,000 people who mostly never saw the product. The same trigger fired on a segment of subscribers who browsed that category in the last 30 days without buying reaches a fraction of the list with a message that reads like a tip instead of an ad. PushEngage’s own industry-wise study found segmentation doubles click rates across industries, and segment-plus-trigger is where both halves compound.
There is a structural advantage in running this on one platform. When web push, app push, WhatsApp, and live chat each live in separate tools, each tool’s frequency cap thinks the subscriber is fine while the subscriber is actually drowning in four channels of messages. One segmentation engine with one subscriber identity makes a cross-channel cap enforceable: the price-drop fires on web push first, and only the high-value segment that did not click gets the WhatsApp follow-up. Segmentation has become a deliverability requirement in its own right, and we cover that argument in full in our companion post on why push notification segmentation is now a deliverability requirement.
The premium-placement era: what to ship this quarter
The bifurcation has a second half that most push programs have not priced in yet. While Android notification silencing buries promos, Android 17’s Live Updates hand lock-screen and status-bar placement to notifications tracking genuinely live, user-initiated events. Google’s I/O 2026 announcements extended the surface to travel, health, and timer use cases, with a Metric Style template that displays up to three live data points. For commerce, the qualifying events are sitting in your order pipeline already: delivery tracking, order status, a flash-sale countdown showing time left and stock left. Promotional blasts can never qualify; Google’s documentation excludes them by name. Triggered notifications are the only architecture that can grow into this placement, because qualifying events are subscriber-initiated by definition.
The quarter’s work, concretely:
- Audit your last 30 days of sends. Bucket every campaign as blast, segmented blast, or triggered. Blast-first programs typically find the bulk of their volume in the first two buckets and most of their push-attributed revenue in the third.
- Ship the first three trigger campaigns: cart abandonment, browse abandonment, welcome. All three run from templates, not custom builds.
- Set exit conditions and a cross-channel frequency cap everywhere, so growing triggered volume never re-creates the blast pattern.
The retention math is the closing argument. Your subscriber list is an asset you already paid acquisition costs to build. Blasts now spend that asset twice, once in unsubscribes and once in platform reputation. Triggered push notifications compound it: every message lands because a subscriber showed intent, and every click feeds the engagement score that keeps your reach intact.
A store sending 30 daily blasts a month is one flagged week away from watching a sale send trickle out at 1,000 messages per minute. The store that replaced those blasts with six trigger campaigns never thinks about the ladder at all, and its push line item reads in recovered revenue. Making that shift takes an afternoon with templates and a free plan that includes every channel: see PushEngage pricing to start with up to 200 subscribers, triggered campaigns included.
