What I will teach you: How to ensure your web push notifications don’t get flagged or blocked by Chrome’s new spam detection system, using real-world testing, simple language shifts, and smart campaign practices.
Why This Matters: Chrome’s New Spam Detection System
In May 2025, Google Chrome introduced an on-device machine learning system that filters out push notifications it sees as misleading or spammy. If your message triggers this filter, users might never even see it. That means lower engagement, wasted campaigns, and lost sales, even if your notification is legit.
Chrome’s update includes a machine learning model that runs directly on Android devices, powered by Gemini and trained on both real and synthetic data. This enables Chrome to detect suspicious push content locally, making it faster and more private, as it doesn’t rely on sending user data to external servers.
When a notification is flagged, Chrome displays a warning showing the site name and gives users options to unsubscribe or view the message. The model evaluates titles, bodies, and even action button text.
Read the official blog post here: Chrome Fighting Unwanted Notifications
Additionally, Chrome’s Safety Check feature now runs automatically in the background, revoking permissions from sites users no longer visit or that send misleading messages. Users can also unsubscribe from push messages directly from the notification panel, contributing to a 30% reduction in total notifications on supported devices.
This is especially critical if you rely on push for:
-
Flash sales
-
Delivery alerts
-
Limited-time campaigns
-
Re-engagement with inactive users
Why Most People Fail at Push Notifications
Most marketers and developers test once, see it works, and assume all is good.
But Chrome’s system learns. That first message might go through, but the same message tomorrow? Blocked.
Even worse, they use language that sounds great but gets flagged — words like "FREE," "WIN," or vague titles like "Act Now" or "Delivery Alert."
My Real-World Testing Experience: What I Learned
I was helping run a push campaign for a beauty brand, and this is what I learned:
-
Sent: "KLEAN BEAUTY PRODUCTS ARRIVED - FREE Delivery!" — Delivered the first time
-
Sent it again a few hours later — Blocked
-
Removed the word FREE — Worked perfectly
-
Added back FREE — Blocked again
-
Tried title: "Klean Beauty to Cart Delivery!" — Still blocked (too vague)
-
Added a red siren emoji 🚨 to the title — Also blocked
That’s when I realised: this isn't just about what you say, it’s how you say it.
Topical Takeaway: Chrome Is Actively Filtering Your Push Content
Chrome now actively filters your push content based on wording, repetition, and user behaviour. Ignore this, and your campaign performance will tank.
7 Rules to Avoid Getting Flagged by Chrome
-
Avoid buzzwords: FREE, WIN, URGENT, or ACT NOW are red flags.
-
Don’t use vague phrasing: Titles like "to cart delivery" or "click here" confuse users and get flagged.
-
Don’t imitate system messages: Chrome watches for notifications that look like OS/system alerts.
-
Test more than once: A message working once doesn’t mean it’ll work twice.
-
Send to your device first: Repeatedly. If it disappears or shows a warning, change it.
-
Be transparent with opt-ins: Never trick users into allowing notifications.
-
Keep it clear and concise: "Your Klean Beauty order qualifies for delivery" works better than "Klean Beauty to Cart Delivery!"
Bottom Line: Adapt or Get Blocked
If you're sending push notifications in 2025 and not adapting to Chrome's new system, you're flying blind. Learn the rules, test smart, and keep your language honest and clear.
This small shift can mean the difference between a successful campaign and one that never even lands.
That’s it, folks! Hope it was a good read 🚀