A good opt-in rate for web push is one that beats your own baseline for the same prompt style and traffic mix. Industry figures often land in the mid-single digits of unique visitors, while acceptance of the browser prompt itself can be higher — and a hard “Block” permanently ends your chance to ask again on that browser.
Opt-in rate is the health metric for list growth. CTR tells you whether messages work; opt-in tells you whether people trust the ask. When I review a push program, I always check how the prompt is timed before I worry about copy on the campaigns.
What is a good opt-in rate for web push?
For publishers measuring subscribers ÷ unique visitors over a year, OneSignal has publicly cited about 4% as a cross-publisher average (OneSignal on web push retention). Other vendor write-ups commonly place media-site web push opt-in somewhere in the 5–10% band, with geography and prompt design moving the needle a lot.
Those are visitor-level figures. They are not the same as prompt acceptance rate (Allow clicks ÷ prompts shown). Chrome UX Report work summarised by OneSignal put average notification prompt acceptance around 17%, with a high rejection share when prompts appear without context (OneSignal analysis of Chrome UX Report data). Google’s own permissions guidance also notes that prompts shown after a user interaction are allowed far more often than prompts with no prior intent (web.dev permissions best practices).
| Metric | What it measures | Rough public context |
|---|---|---|
| Visitor opt-in rate | New subscribers ÷ unique visitors | ~4% cited as a publisher average (OneSignal); many media sites talk about mid-single digits |
| Prompt acceptance | Allow ÷ prompts shown | Chrome UX aggregates historically near ~17% average acceptance |
| Block / deny rate | Explicit Block on the browser prompt | High when the ask has no context; permanent for that origin |
A “good” rate for you is simply: improving month over month without raising bounce or block rates. A 3% visitor opt-in on a high-bounce blog can be healthier than an 8% rate bought with an instant native prompt that trains people to hit Block.
How do you calculate web push opt-in rate?
Use a clear formula and stick to it.
- Visitor opt-in rate = new push subscribers in the period ÷ unique visitors who could have seen a prompt.
- Prompt acceptance rate = Allow responses ÷ times the browser permission prompt was shown.
- Soft-prompt conversion = clicks on your custom “Enable notifications” UI ÷ soft prompts shown.
Don’t mix them. A soft prompt can convert at 20% while the overall visitor opt-in stays at 4%, because most visitors never see either prompt.
Also exclude bots and known crawlers from the visitor denominator when your analytics tool allows it. Inflated uniques make opt-in look worse than it is.
What hurts opt-in rate the most?
The fastest way to kill opt-in is to fire `Notification.requestPermission()` on first paint. web.dev calls that pattern the worst UX: visitors have no reason to say yes, and a Block is sticky (web.dev Permission UX). MDN likewise advises requesting permission only in response to a user gesture (MDN Using the Notifications API).
Other common problems:
- Asking before the reader has seen any content value.
- Vague copy (“Allow notifications?”) with no topic or benefit.
- Asking on every pageview after a soft dismiss, which feels like nagging.
- Ignoring quieter UI and engagement rules in Chrome that already punish noisy sites.
- Not offering a way out later, so people Block instead of unsubscribing cleanly.
How can you improve web push opt-in rate?
Improve the moment of the ask before you redesign the whole site.
- Delay the ask. Wait for a scroll depth, a second pageview, or a finished article.
- Use a soft prompt first. Explain the benefit in your UI; only then call the browser prompt if they opt in. A “Not now” on your UI does not burn the permission.
- Make the benefit specific. “Get alerts when this story updates” beats “Enable notifications”.
- Match prompt type to traffic. High-bounce blogs often do better with a passive bell; engaged product apps may use a modal after an action.
- Separate mobile and desktop. Acceptance often differs by device; tune delay and copy per surface.
- Protect the permission after signup. Low-CTR spam later can cost you the permission in Chrome’s automatic revocation logic (Google, Oct 2025).
For a broader publisher checklist, see web push best practices for publishers. For setup context, how to add push notifications to a website covers the technical side.
How should you read your opt-in rate without obsessing over averages?
Compare like with like.
- Track soft-prompt conversion and browser acceptance separately.
- Segment by landing page and referrer. Homepages and article pages convert differently.
- Watch Block rate, not only Allow rate. Rising Blocks mean your timing is wrong even if Allows look fine for a week.
- Rebuild the baseline after a major prompt redesign; don’t compare the old prompt’s month to the new one without a note.
If a third-party figure can’t be verified for your niche, omit it from your planning doc and teach the team to baseline their own numbers — the same approach I use in push notification metrics.
Where does LaraPush fit?
LaraPush is a self-hosted web push panel with customisable prompts (native, modal, full screen, restrictive and bell), delay and re-prompt control, and optional removal of the deny button on soft prompts. Analytics cover growth alongside campaign CTR. You install it on your own VPS; a licence covers one server with unlimited domains on that server; there is no free plan (Startup from $499, Pro from $799 one-time). See features, pricing and the setup guide. To compare vendors, start from the best push notification services and tools pillar.
Frequently asked questions
A good rate beats your own baseline for the same prompt and traffic mix. Publisher averages around 4% of unique visitors are sometimes cited, and media sites often talk about mid-single digits, but prompt design and geography matter more than matching a published average.
Opt-in rate usually means new subscribers divided by visitors. Prompt acceptance means Allow clicks divided by browser prompts shown. Soft-prompt conversion is a third metric. Track them separately so you know which step is weak.
Timing, copy and prompt type all change who sees the ask and who Blocks. Compare a full month of the new prompt against a full month of the old one, and watch Block rate as well as Allow rate.
Not with the browser prompt for that origin. The user must change the site permission in browser settings. That is why a soft prompt first is safer: declining your UI does not permanently deny the permission.
No. Aggressive prompts can raise short-term opt-ins while attracting low-intent subscribers who never click and later hurt engagement metrics. Pair opt-in with CTR and unsubscribe rate.




