Analytics Kotlin Firebase Plugin

Firebase is Google's platform for mobile apps. The Segment Firebase destination requires that you bundle the Firebase SDK with your project. The Segment-wrapped destination code then runs on the user's device, and sends its tracking calls to the Firebase API endpoints, and a copy to Segment for archiving.

Firebase's destination plugin code is open source and available on GitHub.

Adding the dependency

To install the Segment-Firebase integration, add this line to your gradle file:

implementation 'com.segment.analytics.kotlin.destinations:firebase:<latest_version>'

Or the following for Kotlin DSL:

implementation('com.segment.analytics.kotlin.destinations:firebase:<latest_version>')

Using the plugin in your app

Open the file where you set up and configure the Analytics-Kotlin library. Add this plugin to the list of imports.

import com.segment.analytics.kotlin.destinations.firebase.FirebaseDestination

Under your Analytics-Kotlin library setup, call analytics.add(plugin = ...) to add an instance of the plugin to the Analytics timeline.

1    analytics = Analytics("<YOUR WRITE KEY>", applicationContext) {
2        this.flushAt = 3
3        this.trackApplicationLifecycleEvents = true
4    }
5    analytics.add(plugin = FirebaseDestination(applicationContext))

Your events will now begin to flow to Firebase in device mode.

Identify

When you call Identify, Segment will map to the corresponding Firebase Analytics calls:

If there is a userId in your Identify call, Segment triggers setUserId using the Firebase SDK.
If there are traits included, Segment will set user properties for each trait you include on the Identify call

You can use these traits to create audiences and views to analyze your users' behavior.

Note: Google prohibits sending PII to Firebase unless "robust notice" is given to your app users. For iOS apps, some Analytics features, such as audiences and campaign attribution, and some user properties, such as Age and Interests, require the AdSupport framework to be enabled.

Learn more about Firebase's reporting dashboard here.

Firebase has strict requirements for User Property names. They must:

Begin with a letter (not a number or symbol, including an underscore)
Contain only alphanumeric characters and underscores
Be no longer than 40 characters

User Property values must be fewer than 100 characters.

You are limited to 25 unique user properties per Firebase Console.

Firebase automatically collects these user properties.

Track

When you call Track Segment will log the event with Firebase. Firebase automatically tracks the events listed here and it will still do so when bundling with Segment.

Firebase has a limit of 500 distinctly named events so it pays off to be intentional in what you track.

When you call Track, Segment maps from the Segment spec to those that match Firebase's spec. For anything that does not match, Segment will pass the event to Firebase as a custom event. Custom parameters cannot be seen directly in the Firebase Analytics dashboard but they can be used as filters in Audiences.

Like with user properties, Segment will perform the following transformations on both your event names and event parameters. Unlike user properties, you don't need to pre-define event parameters in your Firebase dashboard.

Trims leading and trailing whitespace from property names
Replaces spaces with underscores
Trims property names to 40 characters (Android only)

Event parameter values must be fewer than 100 characters.

Event mappings

Segment adheres to Firebase's semantic event specification and maps the following Segment spec-matching events to the corresponding Firebase events:

Segment Event	Firebase Event
Products Searched	`search`
Product List Viewed	`view_item_list`
Product Viewed	`view_item`
Product Clicked	`select_content`
Product Shared	`share`
Product Added	`add_to_cart`
Product Added To Wishlist	`add_to_wishlist`
Checkout Started	`begin_checkout`
Promotion Viewed	`present_offer`
Payment Info Entered	`add_payment_info`
Order Completed	`purchase`
Order Refunded	`purchase_refund`

Property mappings

Segment maps the followed Segment spec-matching properties to the corresponding Firebase event parameters:

Segment Property	Firebase Property	Accepted Value(s)
`category`	`item_category`	(String) "kitchen supplies"
`product_id`	`item_id`	(String) "p1234"
`name`	`item_name`	(String) "Le Creuset pot"
`price`	`price`	(double) 1.0
`quantity`	`quantity`	(long) 1
`query`	`search_term`	(String) "Le Creuset"
`shipping`	`shipping`	(double) 2.0
`tax`	`tax`	(double) 0.5
`total`	`value`	(double) 3.99 or (long) 3.99
`revenue`	`value`	(double) 3.99 or (long) 3.99
`order_id`	`transaction_id`	(String) "o555636"
`currency`	`currency`	(String) "USD"

Passing revenue and currency

Ecommerce events containing "revenue" or "total" must also include the appropriate ISO 4217 "currency" string for revenue data to populate to the Firebase dashboard. If a "currency" value is not included, Segment default to "USD".

1analytics.track("Order Completed", buildJsonObject {
2    put("orderId", "order-123")
3    put("revenue", 23.00)
4    put("currency", "USD")
5});

Screen

Segment doesn't map screen events to Firebase because Firebase's SDK collects screen information out of the box for you.

For iOS, you can configure recordScreenViews which will automatically track screen views, or pass in a screen manually using a screen call. You can disable the Automatic Screen reporting by adding the plist flag FirebaseScreenReportingEnabled to Info.plist and set its value to NO (Boolean).

Google Analytics for Firebase iOS does NOT support the case of manual-only screen reporting. Firebase only supports automatic + manual screen reporting or no screen reporting at all.

Conversion tracking and Adwords conversions

Firebase is Google's recommended method for reporting conversions to Adwords. To use Firebase, track the conversion events as you normally would with Segment and Segment will send them through to Firebase.

Troubleshooting

Firebase has great logging. If you're having any issues, you can enable debug mode as outlined in Google'd Debug view documentation.