• Home
  • Integrations
  • SDKs
  • Guides
  • API docs
    No results for ""


    Experimentation events

    Read time: 3 minutes
    Last edited: Mar 09, 2023


    This topic explains Experimentation events and how they work.

    Understanding Experimentation events

    An Experimentation event is any event an SDK produces that impacts your experimentation event count. There are two categories of Experimentation events:

    • Metric events: click, page view, and custom events from metrics used in running experiments.
    • Evaluation events: feature events from flags with targeting rules included in running experiments.

    Understanding metric events

    All LaunchDarkly SDKs send custom metric events when you call the SDK's track method. To learn more about the track method, read Sending custom events.

    In addition, the JavaScript SDK also sends a click metric event when an end user clicks on a page URL or anchor tag that you have included in a click conversion metric, and sends a page view metric event when an end user visits a URL that you have included in a page view conversion metric.

    You can view metric events on the Live events page using the Experiments filter.

    If you prefer to import metric events from your existing data pipeline, read Importing metric events.

    Understanding evaluation events

    All LaunchDarkly SDKs send a feature conversion event when you call the SDK's variation method for a flag in an experiment. To learn more about the variation method, read Evaluating flags.

    In addition, the JavaScript SDK and JavaScript-based SDKs also send a variation evaluation event when you call the SDK's allFlags method, unless you configure it not to. The React Web SDK automatically calls allFlags when you initialize the SDK, and when a context receives a new variation for a flag. To learn more, read Getting all flags.

    Configuring the the JavaScript SDK and JavaScript-based SDKs not to send evaluation events when you call allFlags can help keep your Experimentation costs down. To learn more about how SDKs send events, read Recording events.

    Managing Experimentation billing costs

    You may be billed for experiment keys instead of events

    Customers on newer contracts are billed by unique experiment context keys, rather than by event count. To learn more, read Experiment pricing. If you are unsure about how you are being billed for Experimentation, contact Sales or your LaunchDarkly Account Executive.

    Every time an end user triggers a metric that is attached to a running experiment, your Experimentation event count increases by one. If you are being billed by event count, you may want to reduce your Experimentation event counts to control billing costs. You can manage your event counts by not calling variation and track methods unnecessarily.

    Decreasing metric events

    To reduce page view and click metric events, first update your metrics to use custom instead of click or page view where possible. Then, reduce custom metric events by limiting the number of times you call track.

    You can also cache when you call track to ensure you call it only once per context.

    Decreasing evaluation events

    To reduce evaluation events, cache or persist the response from the SDK, so that you only call the variation method when when you want to register an impression, or when you don't have the response cached or persisted.

    You can also decrease the number of targets in an experiment using audience allocation. To learn how, read Allocating experiment audiences.

    Correlating variations and conversions

    A conversion is when a user action provides a measurable value for a metric. For example, if a metric is tracking clicks on a checkout button, LaunchDarkly registers a conversion each time an end user clicks the button. The SDK sends a track call to register the conversion in LaunchDarkly.

    LaunchDarkly uses variation calls to correlate the variation an end user received to a conversion within a 90-day window. If there are more than 90 days between the time of a variation call and a track call on a user key, then LaunchDarkly does not register the metric in the experiment.

    To learn more about variation calls, read Evaluating flags. To learn more about track calls, read Sending custom events.

    In rare cases, unexpected behavior may cause an end user to receive multiple variations in the 90 days prior to a conversion. In these cases, LaunchDarkly attributes the conversion to the first variation the end user received in that 90-day period.