If you're looking to understand, edit, or update your S3 event mapping, you're in the right place!

Integration covered in this article

Prerequisites

You are a Pro Plan customer
You connected an S3 bucket to HG Insights
You know what the event mapping is
If you want to make edits, you have the architect or admin role

What is the S3 event mapping?

If you send behavioral data to HG Insights via an S3 bucket, you'll find a S3 event mapping that standardizes your user interaction data to be used across HG Insights, in reports and in your behavioral models.

Where do I find the S3 event mapping?

In app.HG Insights.com > Mapping > Event mapping > Amazon S3

Click on 'Amazon S3' to see how your S3 events are mapped to HG Insights.

Click on the Edit logo to access the detailed configuration.

This mapping is pre-populated for you with a 1:1 mapping. This means that each single event from your S3 bucket is mapped to a HG Insights event with a similar name.

How to read the S3 event mapping page?

The first thing to note is that several versions of your S3 event mapping are stored here.

Click on the list in the top right corner to see:

Your Live event mapping, with its date of publication
Your Draft version, if you start making edits and save them, but don't publish them (publishing is currently not a feature accessible to our users)
Previously published event mapping with their date of publication. These are the events mapping that were once live, but have been replaced.

Storing several versions allows to work on a draft before publishing, or to revert to a previous version easily if an incorrect event mapping was published.

The rest of the page presents one event per line.

Exclude from mapping: If the box is checked, this event is not taken into account for your behavioral scoring (because the event is not a user activity, or the event is redundant)
Event: The name of your event in your system
Negative user activity: Check this box to identify negative behaviors (unsubscribe, email bounce...) so they don't appear in the signals seen by your Sales teams and they're not counted in your aggregations
MK Event name (signals): How the event will appear in your CRM signals. No signal is displayed for excluded events.
MK Event name: How HG Insights identifies this event throughout the platform.
Activity type: Helps you categorize your events. 'App Usage' for Mixpanel events.

Your event mapping might use dynamic values in some of these fields. Example:

Viewed page: app.HG Insights.com{{path}} displays the value of the URL path (i.e. the part after the domain), so that your Sales teams know exactly what app page was viewed, and can personalize their outreach based on it.

Viewed page: app.HG Insights.com{{path}} will display Viewed page: app.HG Insights.com/homepage, Viewed page: app.HG Insights.com/integrations/salesforce, etc.

You might also encounter {{*}}, the wildcard character. It is used to represent one or more unspecified characters, to match multiple variations of a word or phrase. Example:

The event sign_up_with_{{*}} will match the events sign_up_with_email, sign_up_with_okta, sign_up_with_google, etc and map them all into a single "Signed Up for MSI" event.

How to edit the S3 event mapping?

Here are some important questions to answer before you start making any modifications to your event mapping.

What events to include from S3?

Any behavioral data that is not available in another system that directly connects to HG Insights.

What events to exclude from S3?

Non-user activities: This means product or marketing events that happen when your user is passive. They don't show any intent from your users.
Redundant events: If an activity performed by your users is represented by 2 different events in your systems, map one of them and exclude the other in order to not give double weight to this activity.

For which activities should S3 be the source of truth?

If you track the same behavioral data in different systems, which system should you choose as the source of truth for each activity type?

As a general rule, if your behavioral data is stored in a system that directly connects with HG Insights, we recommend using this integration.
App Usage: Depending on the volume of behavioral data you have, we will recommend you to use a product data system (like Segment or Mixpanel) or to use an S3 bucket.
Web visits: We recommend using Marketo as the source of truth for web visits. If you don't track web visits in Marketo, you can use S3 as the source of truth.
Hand-raising activities (Demo requests, Contact us forms...): We recommend using Salesforce Campaigns as the source of truth for hand-raising activities. If you don't track hand-raising activities in your CRM or in Marketo, then Segment can be your source of truth.
Marketing activities: We recommend using Salesforce Campaigns as the source of truth for hand-raising activities. If you don't track hand-raising activities in your CRM or in Marketo, then S3 can be your source of truth.

How granularly should your S3 events be mapped?

We recommend keeping the default granularity of the 1:1 mapping.

How to modify the S3 event mapping?

Now that you know what you're about to modify, here's a guide on how to do it.

1. If you want to start working from your current mapping, stay on the live version.

If you previously started working on a draft and want to pick up where you left off, go to your Draft.

2. Click 'Start draft' or 'Edit draft' in the top right corner.

3. Please regularly save your work to avoid being disconnected from the platform and losing your modifications.

Exclude an event
- Check the box 'Exclude from mapping'

Get less granular: group events together
- You may need to group events to create a more significant volume in your training dataset:
- Name the events you want to group with the same name in column 'MK event name'
- You can leave different labels in column 'MK event name (signals)' if you wish to display more granular information to your Sales team.

Get more granular: add conditions
- You may need to create several MK events from a single Mixpanel event with different properties.
- Click 'Add condition'

- In the pop-up window, click 'Add a new condition'
- Select the property you want
- Select a condition on the value of the property
- Select the value of the property
- Check 'Case insensitive' (it allows to find the value regardless of upper/lower case)
- Click 'Save' at the bottom of the pop-up

Maximize signals granularity
- Display any property from your behavioral data source in the signals dynamically, using dynamic values with the syntax: {{name of the property}}
- Don't use dynamic values for "MK event name", as the goal is to group similar events in the Likelihood To Buy model.

Add an event
- Click 'Add new events' in the top right corner
- A new line appears on top of the others
- Type the name of the event as it appears in your system
- Add any condition on the event properties if you need
- Type a user-friendly name for your Sales team to see in the signals in column 'MK event name (signals)'
- Type a user-friendly name in column 'MK event name'
- Select the Activity type

4. Click 'Save draft'.

5. Your edits are now saved in draft.

6. To deploy your changes to production, click 'Publish changes'.

What's next?

After your new event mapping is deployed, you need to take several steps to have this new event taken into account in your behavioral scoring:

Wait for the map process to run in the process page. This step is needed for the platform to take into account the changes published to the event mapping.
Head to the Studio. Duplicate your live LTB model.
Re-build the LTB dataset in the Data Studio. This step is needed to recompute all the data analysis and give you fresh stats and insights! Click 'Change dataset' on the Overview page
Don't change any parameter, just click 'Build and load dataset'
Wait until you receive a success email saying your dataset has been loaded.
Head to Model > Insights to see your new event.
Head to Model > Model to set a weight for this new event.