Impact on performance
This SDK mainly receives critical fixes. New capabilities are shipping in the CSQ SDK first.
Contentsquare strives to be non-intrusive and transparent to client app developers. The following metrics help you evaluate the performance impact of the Contentsquare SDK on your apps.
Performance impact
Section titled Performance impactThe following are the technical specifics we can share on performance.
The performance results were obtained under the following conditions:
| Condition | Value |
|---|---|
| Device model | Pixel 5 |
| Android version | 34 (Android 14) |
| Property | Value |
|---|---|
| SDK size | 1567 KB |
| Nbr. of methods | 5753 |
| Nbr. of fields | 3701 |
| SDK CPU peak on event | <15 % |
| Data transmitted over network for a default batch size of 50 events. The size of the batch can be customized if needed. | ~25kb |
Startup time
Section titled Startup timeStartup time is the additional duration (in milliseconds) the SDK adds when your app starts. It is measured in automated nightly tests on physical devices across two hardware generations: a recent flagship (Google Pixel 9, 2024) and an older generation device (Google Pixel 6, 2021). Together, these measurements demonstrate how performance impact varies across hardware generations and help you estimate expected performance on your users' devices.
| Device | Android version | Median added startup time | Observed range |
|---|---|---|---|
| Google Pixel 9 (2024) | Android 16 | ~7.3 ms | 6.7–9.3 ms |
| Google Pixel 6 (2021) | Android 12 | ~22.7 ms | 18.0–24.3 ms |
Figures were measured using CS SDK for Android version 4.53.0. Results for other SDK versions may differ.
Devices with less processing power than the Pixel 6 may experience higher startup impact than the values shown above.
Startup time depends on several factors:
- The complexity of your app and its initial setup tasks
- The device model, processing power, and system load (such as active background apps, battery level, or thermal throttling)
- The Android version
- When the SDK is initialized and which SDK features are enabled
To measure startup time in your own app, test on the devices and Android versions most relevant to your users. Compare two builds that differ only by the SDK integration, running them on the same device under identical conditions. Repeat each measurement multiple times and compare median values rather than single runs. You can use the Android Macrobenchmark library ↗ with StartupTimingMetric.
Performance test procedure
Section titled Performance test procedure| CPU Usage Test | RAM Usage Test | Memory Leak Test |
|---|---|---|
| We are profiling our Demo app with and without the SDK dependency by following the same use case. | We are profiling our Demo app with and without the SDK dependency by following the same use case. | We are profiling our demo app and we are using the ADB Monkey tool to create random user input events |
| We are measuring the CPU peak for both versions of the app and we are computing the difference | We are measuring the RAM usage peak for both versions of the app and we are computing the difference | During this time we are monitoring the RAM behavior for any anomaly. We are using the Memory Profiler to track/fix the issue |