Thanks!

Predictive analytics

Know what to do, right from the start

Our predictive analytics solution gives you clear and actionable insights into your customers’ LTV so you can confidently optimize your campaigns, and predict long term results, without having to wait

Privacy-centric insights that help you make good choices fast

Make smarter decisions faster with machine learning

The more time you spend waiting for campaign results, the more money you may be wasting.
Our predictive analytics engine uses machine learning to identify and map correlations between early user engagement indicators and expected user value to give you accurate predictive insights early in a campaign’s lifetime.

Put your customers’ privacy at the center

When you respect user privacy you build stronger, longer lasting relationships.
Predictive analytics give you the ability to view aggregate customer behavior indicators without relying on user identity, so that you can base your decisions on comprehensive engagement data while preserving customer privacy.

Make data-driven optimization decisions

More data enables better decisions.
Our predictive analytics engine gives you accurate, predictive KPI values based on every measurable event, within a minimal time frame, so that you can make all of your post-install engagement data count when making optimization decisions.

Key features

An ever evolving industry requires forward thinking solutions. Discover how our predictive analytics solution can help you optimize your campaigns, without the wait, while preserving customer privacy.

Predictive analytics dashboard

Get clear data science insights for your marketing efforts through one comprehensive dashboard

Seamless SKAdNetwork communication

Sync with Apple’s SKAdNetwork to easily translate predictive KPI values into conversion values

Detailed campaign view

Get a detailed, aggregate view of predicted campaign performance values, broken down to predictive ROAS, ARPU, retention rate, cost and more.

Privacy by design

Preserve your customers’ privacy by basing your decisions on predictive insights and behavioral indicators without relying on user identity.

Personalized predictive model

Leverage the competitive edge of a unique predictive model made for your app’s distinct LTV logic, based on first party historical data and customer behavior correlations.

Media partner postback & API

Give media partners immediate access to predicted KPI values for accurate media quality and performance feedback.

Advanced data science made available to every app developer

Predictive analytics puts the power of AI, machine learning, and advanced data science techniques in your hands, to give you a competitive advantage in an ever-shifting environment.

Frequently asked questions

How can Predictive analytics help with SKAN?

AppsFlyer’s SDK measures all of a user’s engagement data in their initial 24 hours post-install, this information is analyzed by the PredictSK AI engine to answer key user behavior questions:

  • How much money is the user expected to spend?
  • Is this a paying user?
  • What is the predicted retention rate for this user? 
The answers to these questions will be delivered under specific value ranges to create an anonymous user persona, which will be delivered to SKAN using a standard conversion value in accordance with iOS’s six bit requirement.
SKAN will then send the aggregated users’ data alongside campaign and media partner information – giving you an aggregate KPI view for a cohort of users who share a similar user persona under a specific campaign.

Is PredictSK aligned with Apple’s user-privacy requirements?

Yes, AppsFlyer’s predictive analytics solution is fully aligned with Apple’s ATT and user-privacy requirements, and communicates seamlessly with SKAN.

Are all six bits being used when working with PredictSK??

No, five of the six bits will indeed be used for delivering the predicted user KPI value to SKAN.
The additional sixth bit can be used through AppsFlyer’s conversion studio to generate additional measurement indications for user behavior and value.

How soon can I start using PredictSK?

PredictSK is currently in beta, and only available for qualified AppsFlyer customers. Prior to onboarding PredictSK, a machine training period is required in which the predictive analytics AI engine will review your app’s historical data, identify correlations between user engagement points and their eventual value to the app, and translate these into a unique predictive model. App owners are encouraged to add business inputs in the form of highlighting key events, event value association and more, to help the model better reflect their business objectives.

Ready to know what to do right from the start?