OpenBack Features: DeviceDecisions Machine Learning Signal
At OpenBack, we dedicate ourselves to helping you send push notifications at the right moment, when it’s most beneficial for your user. Rather than sending push notifications at times when they will most likely not be engaged with – such as when they’re at work, sleeping, or commuting – our philosophy is to send notifications when the user is relaxing and can spare their attention. The tools we offer on our platform to help you determine when this “perfect moment” is are called Signals. Unique among these is our DeviceDecisions Machine Learning Signal, which uses an intelligent algorithm to learn from both users’ past on-device behaviors and real-time contextual factors.
To learn about how timing and sending notifications at the right moment can boost your click-through rate, read our blog post: Notifications Click-Through Rate Boosted by “Adaptive Scheduling”
What Does the DeviceDecisions Machine Learning Signal Do?
OpenBack has 40+ data Signals that can alert the developer of certain user actions, device qualities, or even external conditions that may be useful for determining the perfect moment for notification delivery. These include Signals for user location prediction, battery power, whether the lockscreen is engaged, whether the headphone jack is in use, etc. The OpenBack SDK on the device is learning in real-time a user’s pattern of behaviors. It is constantly factoring in device data to update its algorithm and optimize notification delivery.
Our newly enhanced DeviceDecisions Signal combines both present real-time contextual factors and previous patterns of on-device behavior to determine that absolute perfect moment. For example, this Signal can determine that, because the user takes a break for lunch at 1:05 pm every weekday, 12:50 pm is the best time to send a push notification offering a meal deal at their local Subway. What’s more, the DeviceDecisions Signal can keep you from sending a push notification at a bad time. (e.g. when the device’s battery is very low.)
The DD-ML Signal is one of the most unique tools in your OpenBack toolbox because of how it builds on algorithms with new information learned. For example, say a mobile app user opened a push notification sent at 9 pm last night. A static algorithm would assume that this means 9 pm is the optimal time to receive push notifications, and it would send a notification every night at 9. However, OpenBack’s dynamic algorith, informed by the DeviceDecisions Machine Learning Signal, would also take into account real-time factors that may affect whether the user would pay attention to a notification sent at 9.
For example, perhaps on Wednesday nights they go to a night class. In this case, a notification that arrives when they’re making a late dinner and getting ready for work the next morning would certainly be unwelcome. Perhaps on Thursday night they forget to charge their phone. In this case, they probably wouldn’t bother opening a notification with a deep link. The DD-ML Signal allows your push campaign to be more adaptive and personable on an individual basis.
Setting Up the DeviceDecisions Machine Learning Signal
There are options to weight the Signals algorithm to optimize it for Interactions, Clicks, or Goal completions. If the developer selects multiple optimization options, it will optimize the delivery moment for all of the options. This signal can boost click-through rates (CTR) by x2, but studies of past campaigns have shown results of as high as x10 CTR.
The DD-ML Signal is particularly useful for re-engagement of lapsed users. Perhaps someone who may need conditions to be absolutely perfect before they will open your app again. This Signal is also useful for marketing messaging campaigns, where it’s crucial to engage with the user at the right moment, when they have an opporunity to engage with your app and message.
For more information about the DeviceDecisions Machine Learning Signal, and how to use it to optimize your push notification campaign, get in touch with one of our experts.