Fred Wilson writes a great post about the potential for mobile notifications (which ironically I noticed from my mobile notification bar on my Android phone).
It’s been 18 months since we sent our first push notifications for Tap Tap Revenge 2 on iOS (then iPhone OS 3.0). And over those 18 months we’ve sent over 2.3B notifications to over 100MM devices on iOS, Android and BlackBerry and we’ve learned quite a bit about what it takes to be successful with mobile notifications.
The first use of notifications we saw from publishers were them looking to reach out to their users and drive engagement. They weren’t pushing, but rather, trying to shove their users back into the app. Unfortunately this doesn’t work like traditional forms of notifications (phone, SMS, email, etc). Your mobile phone is an intent driven device. You pick it up to take action: make a call, check your mail, look for notifications. This means you can’t just blast away at users with your messages and expect them to be delighted at your interruption. You have to start a conversation with them and think more about how to send the notification at the right time, to the right user when they are ready to see it–all with the right content.
If you do blast users with notifications they will silence you, or worse, delete you from their phone. We’ve seen some of our customers learn the hard way from this. When a user disables notifications they’re turning off the life blood of your app; engagement. Given the intent-driven nature of mobile devices you have to carry on relevant and timely conversations with users. (The only caveat here being sports notifications; sports fans have an insanely high tolerance for notifications–like in the hundreds per day. But again, this is relevant content to them which explains their high tolerance.)
Dictionary.com is a great example of Push done right with their “Word of the Day” notifications. Those daily updates are a natural extension of their app and brand, and a great way to keep the app top-of-mind. Other examples include Mashable sending out notifications on trending stories (instead of say every single story for the site), Groupon or Living Social sending out the daily deal, Tapulous doing in-game challenges or opt-in notifications for opt-in content. All of these are about creating a relationship and slowly dripping relevant content to them over time.
Another potential example came up during my recent conversation with Rob Woodbridge over on Untether.tv. So many app developers are still trying to replicate the functionality of their website on mobile. That’s missing the point. Take the Toyota iPhone app. You can fully configure your new Toyota, then get a quote for your vehicle all within the app. Now, call me crazy but I can do that on my laptop, and with a better experience. The app doesn’t add anything to the conversation, nor takes into account how a user may want to interact when their phone is the only computer near them. However, with your app in the customer’s hand, you could create value and conversation with that user. What if instead you could enter your VIN and have detailed information about your car? What if you could tie that to the tune-up and oil change records (instead of how most people track via receipts in their glovebox)? Now, hook in mobile notifications and you’ve got the ability to tell the user that their car needs an oil change and oh, by the way, there is a service bay that is open just a few blocks from you right now. Stop by in the next 20 minutes and we’ll give you 15% off. Now that would be an effective conversation with a user of a mobile app coupled with notifications.
Where we are really starting to see interesting possibilities are when look at a large volume of notifications. We’re starting to see trends across apps on when users consume messages, on the types of messages that users respond to effectively, and most importantly being able to aggregate this data anonymously for use by all of our customers. We’re combing through mountains of data to help make sense of it all. Imagine if Groupon could know the best time of day (for that specific user on that specific day) to send a notification because of anonymous data aggregated from how other users use other apps? This is where we see mobile notifications going along with much more sophisticated tools for publishers to send and understand their impact on usage of their apps.
The challenges are many. Each of these mobile platforms behaves differently and there is anything but feature parity across them. Being able to understand the effectiveness of notifications between iOS, Android, BlackBerry and now Windows Phone 7 is only going to get more difficult. Having a view across all of them will be essential.
We are in the early phases of mobile and just like we saw with the early days of the web, people are trying to apply traditional models to a new medium. Just like print ads didn’t translate directly to the web, the same will be true with mobile and notifications. Learning to harness behavioral data in aggregate will be the real magic I think that will unlock the potential for mobile as a successful channel for publishers to reach users effectively.