Context SDK
Introducing the most intelligent way to know when and how to monetize your user.
Today, whether your app is opened when your user is taking the bus to work, in bed about to go to sleep, or when out for drinks with friends, your product experience is the same. However, apps of the future will adjust into the context of their users' environment.
As app usage has exploded over the past decade, personalization based on user context is becoming increasingly important to grow and retain your userbase. Context SDK enables you to create intelligent products that adapt to users' preferences and needs, all while preserving the user's privacy and battery life using only on-device processing.
Context SDK leverages machine learning to make optimized suggestions when to upsell an in-app purchase, what type of ad and dynamic copy to display, or predict what a user is about to do in your app, and dynamically change the product flows to best fit their current situation.
Commute on the train
Alone and bored at night
In a loud bar with friends
Your users have different needs based on the context of what they are doing and where they are. Shouldn't your app be more personalized to better serve them?
Context SDK takes hundreds of signals and builds a highly accurate and complex model, to identify correlations between a user's activities and their in-app conversion rates.
For example, the above graph shows for how people between 3pm and midnight with a battery level below 60% have significantly lower conversion rates (1.9%) than the rest (2.4%). The X axis represents the time of the day, where 0 is midnight, 0.5 is noon, and 1 is midnight again. The outer bar plot shows the bucket size of the data used.
Above you can see how various signals affects in-app upsell conversion rates. For example:
- Users are 94% less likely to convert when they're running
- Users are 16% less likely to convert when they have audio playing
- Users are 10% more likely to convert when they're sitting on a couch
Context SDK performance
Meta has published data on how "less is more" when it comes to notifications and user prompts: Even though in the short-term, showing prompts on every possible occasion will increase your chances of the user engaging, in the long-run, you are better off showing them only when the user is most likely to convert.
With Context SDK, you can significantly reduce the number of prompts you show to your users, and as a result increase your conversion rates.
The above graph shows the conversion rates of 2 buckets for 6 different apps already using Context SDK. Below an explanation on for App 1:
- Conversion rate of all events where the Context SDK has predicted "Bad" → 2.13%
- Conversion rate of all events where the Context SDK has predicted "Good" → 3.12%
- Resulting in a relative difference in conversion rate between those buckets of +46%
Each app is unique, and the Context SDK will cause a different lift depending on various factors. On the 6 apps where we have enough performance data we observed a relative difference between the two buckets ranging from +36% to +46%.
Context SDK privacy
All processing happens on-device, the SDK doesn't require any additional app permissions, and it will never send any network requests. For more details, check out our privacy policy.
Tech Details
Context SDK leverages the latest Apple CoreML technology, and uses minimal CPU and battery power. The code is entirely written in native Swift, while some of the numerical calculations are done using Objective-C Arrays for even better performance. The SDK uses less than a megabyte of space, including all pre-trained models.
Context SDK is super easy to integrate, e.g. to check if now is a good time to upsell:
Within 2 seconds, Context SDK can accurately identify whether you're walking, sitting, driving, biking, or even taking public transit. We provide a pre-trained CoreML model that was trained on dozens of hours of activity recordings of people of all age groups, body types and heights for optimal accuracy.
There are many more powerful ways to use the Context SDK, sign up below to access to the docs!
Other Insights
We've processed over a million events over the last few months, and want to give you some inspiration on what kind of conclusions and insights you can get by using Context SDK.
Those insights are app-specific, meaning every app will have different signals to look out for. We recommend installing Context SDK in read-only mode to collect the first insights, and to allow us to train a custom model, tailored for your app.
Have you ever considered your app may be used 28% of the time while the user is in their bed? Most of us don't really think about in what kinds of situations our apps are running.
There are many obvious things to consider, like the day of the week, and the time of the day
Distracted users behave differently: Users who have audio playing (music, audiobooks, podcasts) convert 18% less. Don't bother them with unnecessary prompts.
The activity detection happens completely on-device within less than 2 seconds. Context SDK comes with pre-trained models trained on dozens of hours of activity data generated by people of various age groups and body types. Every person walks and sits differently.
Besides the activity, you should consider other signals as well:
- Users in low power mode convert 37% less
- Users with their phone in critical thermal state convert 27% less
- Users on a phone call convert 26% less
- Many more, depending on your app, and your userbase
Context matters, and we have the data to prove it. Large tech companies are already using those techniques to optimise their apps, and now is your chance to benefit from it as well. Sign up below to get started!