How to analyze your User, Event and Account Attribute SchemaUpdated 20/05/2019

Learn to use the Attributes view, hide and show attributes, analyze your data flows and optimize it, see what Hull discovered about your data.

Attribute Schema

As you build and maintain your organization data flow connecting different data sources it becomes crucial to get good understanding of what kind of data Hull ingest.

The Attribute and Event Schemas expose valuable information about the Schema that Hull detected for your Users and Accounts. It is a very useful debugging and exploration tool.

You can access them by clicking "Attributes" and "Events" in the secondary navigation of the "Users" and "Accounts" section of your dashboard.

Here's some reasons you will want to use them.

Screenshot 2019-05-21 at 16.36.36

See if an attribute is being collected

View the list of all detected Attributes

The User and Accounts attribute list displays all the attributes that were detected while ingesting external data. you can sort it, or filter it using the filter bar on the top right.

See how frequently an attribute is being updated

View the number of updates for a given attributes

The second column of the list shows the number of updates for this attribute in the last 24 hours, across all your users / accounts. A very high nubmer can either be normal (because you have some frequently changing data) or an indicator of something wasteful. The other colums can help understand which it is better.

Identify data consistency issues with constant updates

Number of Users affected and Updates/User

The third columm shows for how many distinct users this attribute changed in the last 24h. If you have a very high number of updates but a low number of Users (as materialized by the "Updates/User" counter, then it probably means those users are "flip-flopping" between values or that they are continuously updated. This is a very wasteful pattern that consumes traffic in Hull but also in your external services. It should be solved as it probably creates backlogs and delays in updating data, but also might create unstable values in your destination services.

See if an attribute is being used inside some segments

This column shows the number of segments that rely on this attribute. You can hover to see their names. This data is useful to know if you're still using an attribute of if it's a legacy one. Keep in mind that this is only showing usage in Segments - We don't show here if you're sending and/or using this attribute in external services as there is no way to tell for some of them.

See which source(s) is(are) updating this attribute

This column shows the connectors that generated those attribute changes, and when they did so last. If you see a single attribute being updated by several connectors, you need to apply judgement to see if it's expected or if you're overwriting data from one source with data from another.

Connectors generally try to write in their own attribute groups (i.e. Hubspot captures attributes in the hubspot group) but some don't have these conventions (think SQL connector, Incoming Webhooks etc...)

Mark an attribute as Hidden, activate or de-activate tracking it's value changes

This last column (only visible for User attributes) lets you define the behaviour of the platform when this attribute is updated. You can choose between Hidden, Visible, Visible + Tracked

Screenshot 2019-05-21 at 16.45.27

  • Hidden: The attribute will not be visible in the User List, in User Profiles, or available for use in connectors. Set an attribute to hidden when you want to Delete an attribute. As it turns out, you can't really delete attributes since any update will show it again. As a result, you simply hide it.
  • Visible: The regular status. Visible everywhere
  • Visible + Tracked: This tells the platform to record the changes to this attribute for every user as an Attribute Change event. It's very useful to be able to access and segment on the past values of an attribute (Think: "All users who went from <$500 MRR to >500MRR in the past 30 days) but it consumes an additional Incoming Record for each generated event. So activate only for the attributes you need to track

The attributes detail view

Screenshot 2019-05-21 at 16.51.32

The attributes detail view shows you even more precise information about each attribute:

  • Which segments use it
  • The recent changes for this attribute, with the previous, current value and the affected User
  • The volume of attribute changes over time

This view helps you get a sense of which user(s) are affected by too many udpates for instance.

Event Schema

As User Events are crucial data object next to actual User Attributes there is a separate tab in the Users section of the dashboard. After clicking on the Events tab a list of events tracked recently will be shown.

Screenshot 2019-11-07 at 12.46.19

The list shows following columns:

  • User Event Name
  • Count - total count of events tracked in selected time window
  • Users - total number of users with that event tracked in selected time window
  • Events/User - average number of events tracked per one user in selected time window
  • Connectors - connectors and platforms which tracked that event in selected time window

By default table is sorted by the total count of events. This view helps to understand which events are resulting in the biggest volume of the traffic within your organization.