▸Example Segment (Twilio) call (the "before")
import com.segment.analytics.kotlin.android.Analytics
val analytics = Analytics("YOUR_WRITE_KEY", context) {
collectDeviceId = true
flushAt = 20
}
analytics.identify(userId, traitsOf("email" to email, "plan" to "pro"))
analytics.track("Paywall Purchase", buildJsonObject {
put("value", price)
put("currency", "USD")
})App-binary size affects download conversion (especially on lower-end devices and cellular networks) and storage pressure for users with full phones. Heavy analytics SDKs (Firebase Analytics, AppsFlyer, Branch) routinely add several MB to the binary; Respectlytics's per-platform SDK is under 100 KB.
☑Remove Segment (Twilio) cleanly
-
1
Remove the Segment Analytics SDK from your build (
Analytics-Swift/analytics-android/@segment/analytics-react-native/segment_analytics_flutter) -
2
Remove
Analytics.client(writeKey: ...)andanalytics.track(...)call sites — replace withRespectlytics.track("event_name") -
3
Critically: review your Segment destinations and decide which destinations you still need data flowing to from Respectlytics (most don't — that's the point)
-
4
Delete the Segment workspace's mobile source once events have stopped flowing
-
5
Audit and remove the downstream destination SDKs that Segment was the only reason to forward to (e.g., Facebook Pixel, Google Ads)
⇋Segment (Twilio) vs Respectlytics — smaller app binary
| Segment (Twilio) | Respectlytics | |
|---|---|---|
| Typical contribution to IPA size | — see tool note above | < 100 KB |
| Typical contribution to APK / AAB size | — see tool note above | < 100 KB |
| Transitive dependencies | — see tool note above | None (zero third-party) |
❓Frequently asked questions
How much do these SDKs typically add to my binary?
Varies by SDK. Rough order of magnitude on Android (release AAB): Firebase Analytics ~3-4 MB (with transitive Google Play Services), AppsFlyer ~1-2 MB, Branch ~1.5 MB, Mixpanel ~1 MB. Our [SDK Bundle-Size Comparator](/tools/sdk-bundle-size-comparator/) has up-to-date numbers per SDK.
Does binary size actually affect downloads?
Yes — particularly past Apple's 200 MB cellular-download cap and on lower-end Android devices with limited storage. Google's Play Console has internal metrics showing measurable install-rate differences when AAB size crosses common thresholds (50 MB, 100 MB).
How does the size reduction work in practice?
Removing a single ~3 MB SDK from your release build shrinks the binary by roughly that amount. Multiplied by user installs, the cellular-data savings compound. iOS App Thinning handles per-device variants; Android AAB delivers size-optimised installs by ABI / density / language.
Does the smaller SDK come at a feature cost?
Respectlytics intentionally doesn't bundle features it considers separate concerns: no ads, no attribution, no in-app messaging, no A/B testing, no crash reporting. Each of those is a dedicated tool's job. The smaller surface is the deliberate design — you add only what you actually use.