Why Aren't More People Talking About Push Notification Reliability?
Unreliable delivery is a painpoint that's keeping push notifications from achieving their full potential as a next-gen marketing tool. What is the root of poor reliability? How does it hurt push notifications, and how can we improve on traditional push technology?
In the 21st century, we have accepted that mobile phones are most people’s primary interface with the outside world. We use them to communicate with each other, to buy things, to check the weather/news/football, to stream music and videos, to play games… the list goes on and on.
As of February 2021, 5.25 billion people worldwide own some kind of mobile device. And that number is set to grow to 7.33 billion by 2023. With numbers like these, it goes without saying that the most efficient way for a brand to get its message out to consumers is to send it straight to their devices.
The means of conveying that message are, of course, push notifications. But mobile push notifications are still a relatively new technology, going through growing pains. Currently, the benchmark click-through rate (CTR) for push notifications is abysmally low at 2.74%. This can be boiled down to failures on two fronts: lack of personalization and unreliable delivery.
Personalization is easily fixed on the marketing end. However, reliability is a trickier beast, as it is rooted in the very architecture of push notification software. And to offer a solution to the challenge of unreliable delivery, we must break down how push notifications are sent.
But first, what exactly do we mean by “reliable delivery”? And why is it such a problem in the mobile engagement industry?

Reliable Delivery: A Definition
Email platforms are constantly talking about the deliverability of emails. According to Omnisend, 95% delivery is considered a good delivery rate, with a less than 3% bounce rate. Given how marketing emails need to navigate different spam filters and promotions folders before they reach the user’s attention, one would think push notifications have a much easier job at engaging users’ attention. But this is not the case.
Push notification delivery rates are so unreliable that it makes it virtually impossible to guarantee that a message designed to be read at a specific time – for example, a notification with a promotion for a new sandwich, scheduled to deliver to devices at 1 pm – will reach its destination at that time.
And this closes the door on a lot of possibilities, because one of the unique strengths of push technology as a marketing tool is its ability to reach consumers at the exact moment that they’re most primed to engage with your message.
Why Don’t Some Push Notifications Get Delivered?
Push notification delivery failure is catastrophic to UX, and can drive app users to turn off their notifications, or even delete your app in extreme cases. We’ve all heard horror stories of a series of notifications all hitting the user’s device at once, sometimes in a jumbled order. Or even notifications with crucial information – say, a gate change for a user’s flight booked through the app – failing to deliver entirely. But why is this happening?
A lot of explanations for push notification delivery failure is just a matter of people trying to pass the buck. With standard push notification SDKs, there are a variety of factors that can affect deliverability. Apps that have been recently opened tend to have higher notification delivery rates.
Delivery can also be adversely affected by security firewalls, or by manufacturers that build into the device a function that shuts down non-essential background programs in order to save battery life.
But to get at the real heart of deliverability problems, you have to break down the very architecture of the traditional push notification platform.
.png?width=734&height=413)
The Standard 3-Prong Push Notification Triangle
The technology for sending mobile push notifications has not changed since its inception in 2009, and that’s where the problem lies. With the traditional SDK, user data is gathered from the device in the form of a push token and sent to the app backend server. The push token together with the content of the notification is then sent to the cloud server: Firebase Cloud Messaging (FCM) for Android, or APNS for iOS.
Notifications are stored and processed here, along with billions of other data points, until the cloud server receives a signal from the app, which triggers delivery to individual devices. In order for this to work, the radio network in the app needs to be on constantly to check for notifications in the cloud. This is a leading culprit in battery drain of devices. Once the cloud receives this signal, it blasts out notifications en masse.
There is clearly much in this set-up that can go wrong. For one thing, if the push token changes on the app’s end, the push token will have to re-register with the cloud, which often causes delays. Plus, the cloud server sends out batches of notifications, with the last push token into the server being the first notification sent out.
This raises the risk that the server might overlook earlier messages, and only collect the most recent ones. The entire process of outsourcing data processing to cloud servers is inefficient and often leads to poor UX.
Mobile Edge Computing and a New Era for Push Notifications
Thanks to the advent of mobile edge computing, in which the leveraging of user data takes place entirely on the device itself, sending notifications through the cloud is on its way out the door. Changing to a process where data processing occurs device-side not only fixes reliability failures, but it is also compliant with global data privacy regulations such as GDPR, HIPAA, and COPPA.
What’s more, it allows for an unprecedented level of insights on when and how users interact with your notification, what goal conversions users complete because of clicking on a notification, and much more.
Tracking metrics and long-term trends of user engagement with your push notifications is a possibility that mobile marketers couldn’t even dream of when traditional push technology was rolled out in 2009. And this is what will help us improve our campaigns, learn from our mistakes, and optimize the push notification experience for users.
Receiving reliable notifications at the right moment means users will be more primed to click on your message. Higher CTR will result in greater interaction with your app and, by extension, higher revenues.
Push notifications have so much dormant potential… why should developers throw all that away (and leave revenue on the table to boot!) just because of outdated technology?