Tiered Events

mParticle collects data from all your platforms and data connections and connects it to all of your team’s tools and systems. All collected data is used to resolve identities and stitch together customer profiles and is available to be forwarded to downstream systems. Data can also be stored in mParticle, in case you need to run data replays (re-ingesting data with help from mParticle Support). Additionally, mParticle makes all data available for evaluation in Audiences and Calculated Attributes. However, many events and the data related to them may not be needed for running data replays or evaluations depending on how you use Audiences or Calculated Attributes. With Tiered Events, you can choose to collect data but not store it, or to collect and store data but not evaluate it.

These configurations help improve performance for all customers, and may provide cost savings for value-based pricing customers.

mParticle provides three event tiers. All three tiers support ingestion and forwarding of data, but provide different levels of support for storage and evaluation:

• Connect: Data is not stored in mParticle, but can be ingested and forwarded. The data is also used for profile enrichment. Because Connect tier events aren’t stored, they won’t be visible in the User Activity view.
• Preserve: Events marked as Preserve are ingested, stored, and forwarded but are not used in real-time evaluations for audiences or calculated attributes. The data is still available for replaying to third party tools, and for backfills of calculated attributes and standard audience creation.
• Personalize: This tier is the default for all events and related data collected as part of an mParticle ingestion. Events are ingested, available for evaluation in audiences and calculated attributes, stored, and forwarded.

The following table can help you decide which tier to apply to data:

You can always change an event tier. If your account uses value-based pricing, you can also change the default event tier for ingested events.

Workflow

To change events from the default Personalize event tier, plan and implement the changes:

1. In Data Master > Catalog, click Download Report to download a CSV file that lists all the events used in calculated attributes and audiences, and shows data volumes for the time range you specify.

   The report shows you approximately how many of each event type have been received from input for the date range you selected. You can use this number, for example, to identify events that you may want to change to another tier, and which will be easy to remove from the calculated attribute or audience definition so that you can change the tier for that event type.

2. Identify the events you don’t plan to store or use for evaluation in audiences or calculated attributes. These are the events you’ll change to Connect. Events that you do plan to store, but don’t need for evaluation can be changed to Preserve.
3. Using the mParticle UI (Data Master > Catalog), set the events you don’t need to evaluate in audiences or calculated attributes to another tier.

Prerequisites

• You must have the Admin and Compliance role to change event tiers for an event type. All other data will behave as if on the Personalize tier.

• Verify that the consequences of changing tiers are acceptable in your environment:

  • Event marked Preserve are never available for real-time evaluation, only for data replay or backfills.
  • Events marked Connect are not available for data replays, backfills or real-time evaluation. These events are not visible in the User Activity
  • Only events, including custom events, screen views, commerce events, and application lifecycle events, can be assigned an event tier. User information can’t be assigned a tier and so is only in the Personalize tier.

Change the event tier

To change the event tier:

1. Navigate to Data Master > Catalog and select the event type such as Custom Events or Screen View.
2. Next to the event, click the Event Tier drop-down and select a new tier.
3. Click Apply.

Change the default event tier

By default, mParticle assigns all event types to the Personalize tier. However, you can change the default assignment to Preserve or Connect if your account uses value-based pricing.

Changing the default tier is helpful. For example, if you create a data plan with many new event types and most of them won’t be needed for real-time evaluations, you could assign them to the Preserve or Connect tier by default.

To change the default event tier for a workspace:

1. In the left-hand navigation of the relevant mParticle workspace, click the org name at the top to display Settings.
2. Click Settings.
3. From the list of workspaces, click a workspace name to display its Edit Workspace page dialog.
4. Open Defaults and use the drop-down in the Event Tiers section to choose a new default tier.
5. Save your changes.

After you set a default tier, all new events of all types seen in mParticle are assigned the new default tier. You can override the default tier assignment of an event type in Data Master > Catalog. For example, you could change the default tier to Connect in the workspace setting and also set the tier for event type MyEvent to Personalize. Then, all new and existing events of type MyEvent are assigned the Personalize tier, while all other new events are given the new default, Connect.

Event Volume Report

The event volume report, accessible from Data Master > Catalog, is a CSV file that lists all the events used in calculated attributes and audiences. It also shows data volumes for the time range you specify.

Use this report for diagnostic purposes—to help you understand what data is being ingested in real-time. This report is a snapshot in time.

The report shows you approximately how many of each event type have been received from input for the date range you selected. You can use this number, for example, to identify events that you may want to change to another tier, and which will be easy to remove from the calculated attribute or audience definition so that you can change the tier for that event type.

The event counts in the Event Volume Report are approximate and don’t include some events, such as those with old timestamps.

Event Volume Report and usage-based pricing

If your contract uses usage-based pricing, expect differences between the Event Volume Report, which is a snapshot in time, and your monthly Credit Usage Report, which is a calculation of billable activity over the billable period. Billable activity takes into account the metering logic that the Event Volume Report doesn’t include. For details, see Usage metering for value-based pricing (VBP).

For example, imagine that you ingest one million events of type my_event on day 1 of your billing period, and that event type is in the Connect tier. Then you change the event tier to Preserve on day 5 of your billing period, and ingest one million more events of type my_event. On day 6, if you download the Event Volume Report, you’ll see two million my_event events in the Preserve tier. However, since those first million events were process as Connect tier events, the billing report will show one million Connect tier and one million Preserve tier events of type my_event.

The two reports are not meant to be compared, as they represent different views of your event data for different purposes:

• The Event Volume Report should be used to make forward-looking decisions such as “I can move these six events to tier Connect.”
• The Credit Usage report is backward-looking. It tells usage-based pricing customers how their monthly charges were calculated.

Differences occur for a number of reasons, including (but not limited to) the following:

• The Event Volume Report includes some events like consent grants and denials that aren’t included in the monthly Credit Usage Report, which details usage-based billing for value-based pricing customers.
• You may run the Event Volume Report before recent batches have completed processing. If those batches are processed after you run the report but before your Credit Usage Report is completed, the number of events will be different.
• If your event data has an event timestamp older than 72 hours compared to the time mParticle receives it, the events won’t be shown in the Event Volume Report, but they will be calculated as part of the Credit Usage Report.