The missing data layer in app growth

For much of the last decade, app growth has been driven by signals marketers could easily see: clicks, installs, sessions, and conversions. But beneath the surface, a far richer layer of data has always existed, one that lives entirely on the device itself. Now, in a privacy-first world shaped by the loss of IDFA, that overlooked layer is becoming one of the most valuable assets in the app economy.

At Business of Apps Berlin 2025, Peggy Anne Salz sat down with Dieter Rappold, CEO and Co-Founder of ContextSDK, to discuss how on-device contextual data is redefining growth, monetisation, and user engagement. The conversation explored why timing may be the most underutilised lever in app growth and how motion and sensor data can unlock intent without compromising privacy.

Why context is having its moment

Dieter describes the current environment as a long-overdue correction. As the industry adjusted to stricter privacy standards, many marketers focused on what had been lost. ContextSDK, by contrast, focused on what was always there.

Smartphones are packed with highly sensitive sensors, generating streams of contextual data around movement, orientation, charging state, battery level, and time of day. This data is freely available on-device, privacy-safe, and incredibly revealing when interpreted correctly.

For Dieter, timing is the key insight. Knowing when a user is most receptive can matter more than knowing who they are.

From device assumptions to real-world behaviour

In the early days, device data was used crudely. A high-end phone might signal affluence. A specific Android model might imply a demographic segment. Today, those assumptions are obsolete.

Instead, ContextSDK looks at users in their natural habitat — how they physically interact with their device throughout the day. By analysing accelerometer and gyroscope data, the ContextSDK platform can identify more than 24 distinct real-world activities with up to 96% accuracy.

That includes whether a user is walking, sitting, lying in bed, relaxing on the sofa, or even whether they are travelling in a combustion-engine vehicle versus an electric one. Each context carries a different level of intent and a different likelihood of engagement or conversion.

Why intent beats identity

One of the most powerful outcomes of contextual data is intent prediction. Dieter shared how simple signals like battery level or charging status can dramatically affect conversion probability though the impact varies by vertical and audience.

In one surprising case, a Gen Z dating app saw no drop in conversion even when battery levels fell below 3%. That insight only surfaced because the data was collected, analysed, and tested.

The lesson is clear: context is not universal. It must be learned, not assumed.

Source: Business of Apps via YouTube

From monetisation to data infrastructure

ContextSDK originally launched with off-the-shelf products like Context Decision and Context Push, designed to help apps optimise monetisation and engagement. Increasingly, however, the company is evolving into a broader data infrastructure provider. Large app businesses are now building their own use cases on top of ContextSDK’s data layer.

One of the most compelling applications is fraud detection. Phones that trigger activity without any corresponding physical movement are often a strong indicator of fraud. By combining motion data with machine learning models, ContextSDK can identify fraudulent behaviour in the very first session, potentially saving millions in wasted user acquisition spend.

A new edge for ad networks

Looking ahead, Dieter outlined how contextual signals could reshape programmatic advertising itself. By surfacing a curated set of “golden signals” from hundreds of available data points, ad networks could dramatically improve bidding performance and targeting accuracy.

In a hyper-competitive and transparent market, even small performance gains can compound across massive spend volumes, creating a meaningful competitive advantage.

Engagement is about timing, not volume

Context doesn’t just inform what to show users, it informs when to connect with them.

Through Context Push, ContextSDK integrates with platforms like Firebase, OneSignal, CleverTap, and Customer.io to deliver notifications based on real-world readiness rather than fixed timestamps.

A phone face-down on a table signals low engagement potential. A user relaxed on the couch, phone unlocked and scrolling, signals the opposite. Delivering messages at the right moment can significantly improve engagement without increasing message volume.

The same logic applies to churn prediction. Shorter sessions, increased usage while in transit, and changes in physical context can all indicate declining intent, long before a user actually cancels.

From curiosity to category-defining layer

When ContextSDK first launched, even its founders weren’t sure how valuable this data would be. As Dieter admitted, it could have quickly proven to be a dead-end or something much bigger.

Today, it is increasingly clear that on-device context represents a missing layer in mobile growth. In a world of rising acquisition costs and fierce competition for retention, ignoring an entire category of first-party data is no longer an option.

Looking ahead

As apps compete not just for installs but for moments of attention, understanding real-world context is becoming essential. The future of app growth won’t be defined solely by who users are, but by what they are doing and when.

Watch the full video to discover all of Dieter’s insights. You can also watch all episodes of App Talks here.