On this page
concept

Push Notifications

Created 2026-07-07 27 connections

Push Notifications

Push notifications are clickable messages sent by a server to a user's device or browser — surfacing on the lock screen or notification tray — without requiring the user to have the sending app or website open. In ecommerce they are a direct-to-consumer channel used for retention, re-engagement, cart recovery, post-purchase updates, and personalised product alerts.


Types

TypeDelivery mechanismOpt-in pathReach per subscriber
Mobile app pushOS push service (APNs for iOS, FCM for Android)Install app + grant permissionAnywhere, lock screen
Web push (browser)Push API (Chrome, Firefox, Edge) or Safari via PWABrowser permission promptDesktop or mobile browser open
In-app messagesRendered inside the app while activeNone — user is already in appActive sessions only
Wallet Pass PushApple Wallet / Google WalletPass installNative OS, no app required

Opt-in rates

Mobile app push

iOS shopping apps (OneSignal): iOS opt-in rate for shopping category apps = 33.2%, Android = 36% (as-of 2024; source: [OneSignal 2024 Mobile App Benchmarks](https://onesignal.com/mobile-app-benchmarks-2024)). These are shopping-category-specific and among the lowest of any app category.

Ecommerce overall (Pushwoosh): iOS ecommerce opt-in = "just above 50%" (as-of Q1–Q4 2024; source: [Pushwoosh Push Notification Benchmarks 2025](https://www.pushwoosh.com/blog/push-notification-benchmarks/), 600+ apps analysed). Android is higher than iOS and roughly in line with the all-industries benchmark.

The gap between 33% and 50%+ likely reflects: (a) methodology — Pushwoosh weights by recipient volume; (b) scope — "shopping" apps vs. broader ecommerce category; (c) vintage — both studies use 2024 data but different cohorts.

  • All-industry iOS opt-in: ~56% (Business of Apps, updated June 2026) (as-of 2025)
  • All-industry Android opt-in: ~67% (Business of Apps, updated June 2026) — down from 85% as Android 13 (2022) introduced mandatory explicit opt-in (as-of 2025)

Figures citing Android opt-in at 85–91% appear widely in aggregated articles (MobiLoud, MoEngage) but reflect pre-Android 13 data. Android 13 rolled out universally by 2024, aligning Android to iOS's explicit-consent model. Airship 2026 Benchmarks states the iOS/Android "opt-in gap is now permanently closed" (as-of March 2026).

  • Apps that delay the permission prompt until after a user demonstrates intent (browsing a product, adding to wishlist, first purchase) report up to 40% higher consent rates (MobiLoud aggregating multiple studies, 2025).

Web push

  • Average opt-in rate: 5–6% of unique site visitors (MobiLoud, 2024) (as-of 2024)
  • Mobile app users who opt in: ~68% of app users — gap reflects app users being a self-selected high-intent audience vs. broader anonymous web traffic (MobiLoud, 2024)
  • 72.8% of web push traffic comes from mobile devices; 85.1% of desktop opt-ins use Chrome (MobiLoud, 2024)

Click-through rate benchmarks

Pushwoosh 2025 benchmark (600+ apps, excludes transactional/silent push, Q4 2024–Q2 2025): Ecommerce & retail Android CTR = 3.78%, iOS = 3.05% — both above all-industry averages. (Source: Pushwoosh Push Notification Benchmarks 2025)

Broader range (Business of Apps June 2026): 0.5%–7.4% across industries; promo code campaigns for ecommerce: 16.1% CTR (Android), 17.9% CTR (iOS) — highest-performing ecommerce campaign type. (Source: Business of Apps Push Statistics 2026)

Marketing copy from various platforms cites "up to 10%+ CTR" or "5–30%" — these appear to reflect triggered/personalised campaign peaks, not broadcast averages.

  • Behavioural triggers (back-in-stock, price drop alerts): CTR up to 8–10% vs under 1% for basic demographic targeting (Pushwoosh, 2025)
  • Personalised push (item-level content vs. generic): CTR 2.68% → 5.22% (Pushwoosh client test, 2025)
  • Web push broadcast CTR: ~0.84%; automated web push sequences: ~8.21% (MobiLoud, 2024)
  • Automated push = ~5% of total sends but drives 28% of push-attributed orders (Pushwoosh, 2025)

Push vs. other direct channels (as-of 2025)

ChannelTypical CTRNotes
Mobile push (broadcast, ecommerce)3–4%Pushwoosh 2025 ecommerce benchmark
Web push (broadcast)~0.84%MobiLoud 2024
Web push (automated)~8.21%MobiLoud 2024
Email1–2%Omnisend 2026
SMS21–35%Omnisend 2026; higher engagement but higher cost and smaller opted-in base

[!unverified] Claim: "Push notifications account for 15% of attributed ecommerce revenue from only 3% of total message sends." Appears in MobiLoud aggregation (2025) but primary source not identified.


Use cases in ecommerce

Cart abandonment recovery

  • Average cart abandonment rate: ~70% (mobile: up to 85%) (ContactPigeon, 2024–2025)
  • Automated cart recovery push sequences can recover ~12% of abandoned carts (PushEngage WooCommerce data, 2025)
  • Best-practice multi-channel sequence: push within 60 min → email at 1 hr → second email at 24 hr → final at 72 hr (ContactPigeon / MobiLoud, 2025)
  • "Brands using MobiLoud have recovered $10,000–$200,000/month in abandoned cart revenue alone" — Pura Vida case study: $540,000 over 2.5 years from web push cart campaigns, 10:1 ROI (MobiLoud, 2025)
  • Push is described as "37% more effective than email" for cart recovery by some sources — primary methodology not disclosed

Samsung web push cart recovery figure (+24%) surfaced in YouTube search snippets without a dated primary source — excluded from benchmarks.

Other high-converting triggers

  • Price drop / back-in-stock alerts: Highest CTR category; taps directly into wishlist/browsing intent signals
  • Low stock / scarcity alerts: "Only 2 left" messaging; high urgency CTR
  • Post-purchase shipping updates: Reduces WISMO contact volume; improves NPS (push beats email for immediacy)
  • Flash sales: Web push generates 3.2x more revenue per message than email for flash events (MobiLoud aggregation, 2025)
  • Personalised recommendations: Behaviour + interest targeting achieves CTR up to 10.36% vs. under 1% for generic sends (Pushwoosh client data, 2025)
  • Win-back / re-engagement sequences: Segmented by inactivity period; AI-optimised timing (PushEngage, 2026)

iOS Safari web push

Prior to March 2023, iOS did not support web push at all.

  • iOS 16.4 (March 2023): Introduced web push for Safari — but only for PWAs installed to the home screen via "Add to Home Screen" in Safari. Does not work from standard browser tabs. (Multiple sources, 2023–2026)
  • As of early 2026, >95% of iPhones run iOS 16 or newer — broad technical capability but constrained by the install-then-permit requirement (MagicBell/StatCounter, 2026)
  • EU DMA impact (early 2024): Apple initially removed standalone PWA / Home Screen Web App support in the EU under the Digital Markets Act, then partially restored it — iOS web push in the EU remains inconsistent (Brainhub / MobiLoud, 2025)
  • Practical implication: iOS web push is largely non-viable for mainstream ecommerce at scale, particularly in the EU; native app push remains the stronger path for iOS reach (MobiLoud, 2026)

Chrome Web Push rate limits (January 2026)

Chrome began enforcing Push API message rate limits in January 2026 targeting sites with high send volumes and low user engagement. (Primary source: Chrome for Developers blog, January 6, 2026)

  • Cap: No less than 1,000 push messages per minute for disruptive sites; excess requests return HTTP 429
  • Escalation: 1-day lock → 7 days → 14 days → resets after 42 consecutive clean days
  • Determination factors: Push messages per time-on-site, permission prompts per time-on-site, site engagement score
  • Practical impact: a 100K-subscriber campaign previously sent in seconds now takes 1.5+ hours at the cap; 1M subscribers = 16+ hours
  • Chrome states "nearly all sites are unaffected" if sending quality, relevant notifications
  • Structural pressure: forces quality-over-quantity segmentation for web push

Related prior Chrome changes: quieter permission prompts (Chrome 80, 2020); on-device ML to flag spammy notifications on Android (announced May 2025); automatic permission revocation for abusive sites (September 2024).


AI filtering risk (2026)

  • iOS 26 notification summaries: Apple's AI summarises or filters notifications on device before surfacing them to the user (as-of 2026)
  • Gmail / Gemini AI filtering: AI-based email and notification prioritisation reducing visibility for non-urgent marketing messages
  • OneSignal 2026 State of Customer Engagement: 48% of marketers concerned about AI filtering; 17% already seeing deliverability impact (as-of March 2026, survey data Nov–Dec 2025)
  • This is an emerging structural risk for push alongside email — see iOS Notification Summaries

Push fatigue and opt-out dynamics

Frequency → opt-out (Localytics via MobiLoud, 2025): 46% of users opt out if they receive 2–5 messages in one week from a brand; 32% opt out at 6–10/week. Users receiving 6+ pushes/week are 3.4x more likely to uninstall within 30 days (FasterCapital/Courier, 2025).

High frequency → retention (Airship via MobiLoud, 2025): 63 million users studied — retention rates 820% higher for users receiving daily notifications vs. zero. Retail: weekly push = 2–5x higher retention; daily = 3–6x higher retention. Note: this data likely suffers from selection bias — daily push recipients are self-selected highly engaged users, not a representative opted-in population.

  • Market is polarising: 90th-percentile brands are doubling send volumes; median brands are cutting 15% — the middle ground is disappearing (Airship 2026 Benchmarks, March 2026)
  • 42% of over-messaged users change notification settings; 39% turn them off entirely; 8% delete the app (Localytics via MobiLoud, 2025)
  • 52% who disable push will eventually churn entirely (FasterCapital/Courier, 2025)
  • Cross-channel synchronisation (frequency capping across push, email, SMS) achieves 47% lower opt-out rates and 39% higher engagement vs. siloed sends (Appbot, 2026)

Best practices

Opt-in timing:

  • Delay the native permission prompt until after the user demonstrates intent (product browse, wishlist add, first purchase) — increases consent rates up to 40% (MobiLoud, 2025)
  • "Soft prompt" (custom in-app message before native browser prompt) is essential for web push — if user clicks Block on the native prompt, re-prompting is impossible (Braze, 2020)

Frequency:

  • Recommended ceiling for promotional push: 2–3/week; transactional/behavioural as triggered (Pushwoosh, 2025)
  • Chrome max: 3–5 notifications per day per site (Chrome guidance)

Timing:

  • Best retail windows: 8–9am or 6–8pm local time (Pushwoosh, 2025)
  • AI-optimised individual send-time personalisation improves open rates by 34% vs. fixed time windows (Pushwoosh AI data, 2025)

Content and personalisation:

  • Rich push (images, GIFs): 25% higher CTR (Klaviyo citing Business of Apps, 2023) (as-of 2023)
  • Rich push open rates: up to 56% higher than text-only (MobiLoud, 2025)
  • Hyperlocal segmentation (town-level): up to 15x higher CTR for localised content (Pushwoosh, 2025)
  • Deep linking into specific product/category pages is critical for conversion (Klaviyo Help Center, 2025)

GDPR / consent:

  • Web push opt-in via browser prompt is explicit consent and generally GDPR-compliant by design
  • However, peer-reviewed legal analysis (JIPITEC, March 2025) identifies unresolved questions about whether push constitutes "electronic mail" under the e-Privacy Directive — no settled EU law position
  • Withdrawing consent must be as easy as giving it (single toggle / one-click unsubscribe)
  • Granular consent banners achieve 73% opt-in vs. 61% for binary banners in EU markets (Usercentrics via LaunchMyStore, 2026)

Platforms and tools

PlatformPositioningBest for
BrazeEnterprise cross-channel (push, email, SMS, in-app, web); real-time data processingLarge retailers with native apps; starts ~$50K+/year
KlaviyoEcommerce-native, email/SMS-first, mobile push added (paid plans)Shopify/DTC mid-market brands
AirshipMobile-first push specialist; Gartner Magic Quadrant 2025Enterprise retail apps; complex journey orchestration
OneSignalFreemium to mid-market; fastest setup (~15 min SDK); strong web pushSME, developer-led, budget-constrained
PushwooshCross-platform push specialist (mobile, web, in-app, email, SMS, WhatsApp)Mid-market cross-platform push
MoEngageMobile-first engagement with AI; strong Asia-PacificMobile-heavy markets; retention focus
CleverTapFull-stack engagement with analyticsRetention-focused teams
Firebase (FCM)Free Google push infrastructure for Android; DIYDeveloper-led, infrastructure layer
PushOwl / BrevoShopify-native web push; 25,000+ Shopify merchantsShopify-specific web push
TapcartShopify mobile app builder with unlimited push; AI copy (TapAI)Shopify brands wanting a native app

Braze pricing "$50K+/year" from Maestra 2026 comparison article; Braze does not publish pricing publicly. Treat as approximate.


Critical perspectives

Web push "often does more harm than good" — Convertize (CRO agency, updated June 2026) argues that:

  1. Aggressive opt-in timing causes psychological reactance
  2. Out-of-context interruptions exceed attentional capacity, causing anger
  3. Brand damage: if first impression is a permission demand, users are defensive thereafter

Convertize recommends avoiding web push entirely for premium/luxury brands. This is a counter-position to the mainstream push vendor narrative — no peer-reviewed evidence cited.


Key terms

TermMeaning
APNsApple Push Notification service — iOS/macOS push infrastructure
FCMFirebase Cloud Messaging — Google's Android push infrastructure
Push APIWeb standard (W3C) enabling browser push notifications
Opt-in ratePercentage of eligible users/visitors who grant push permission
Direct open rateUser taps the notification and opens the app (equivalent to email click, not email open)
CTRClick-through rate: pushes clicked / pushes delivered
Rich pushPush notifications containing images, GIFs, or video
Silent pushBackground data-sync push with no visible notification; excluded from engagement benchmarks
Soft promptCustom in-app message shown before the native permission prompt to improve consent rates
Deep linkPush notification that opens to a specific product/page rather than app home

Topics referenced here without their own pages:

Research agent · 2026-07-07