The Analytics Suite is at a major turning point in its history. A brand-new data engine is under development and you will see the early results in the new Data Query, which is scheduled for beta in October 2019:
- Optimized query performance,
- On-the-Fly Unique Visitors and Unique Identified Visitors,
- Custom metrics and Global Segments (usable on all sites, regardless of properties and metrics used),
- Multi-site analyses without constraints (including segmented),
- Crossing properties without constraint,
- We are keeping a few surprises for the release!
📢 To take full advantage of these new features, you need to optimize the configuration of your site, page and order indicators before the 12th of August 2019 ⏲️. If you have any questions or need assistance, please get in touch with our Support Centre (email@example.com).
Here is a video recap of these new features and the actions you need to take:
If you have any questions or need assistance, please get in touch with our Support Centre (firstname.lastname@example.org).
❓ Can AT Internet help me with this alignment?
Absolutely! Please contact your Support Centre (email@example.com), your Customer Success Manager or your dedicated consultant.
❓ I don't have a lot of time to do this today!
You have until the 12th of August 2019 to rename all the custom variable labels you can to align your data model. Within this time, you will need to validate that the custom variable labels are appropriate for you and that we can use this configuration as a basis for the automatic creation of your new data model.
❓ What are the benefits of this new data model!
❓ When and where will I see the benefits of this new data model?
In the short term, standardizing the labels of your customized variables will simply serve to "clean" your configuration. In Explorer, Dashboards, Reporting API, Data Query and Data Flow, you will see the new custom variable names appear, but the data engine will always be the same.
The real benefit of this new data model will be tangible when the beta of Data Query 3 is released in October 2019 because it will benefit from this new engine. Gradually, Explorer and the other interfaces will use this new data engine.
❓ What happens if I don’t do anything?
You will not lose any of your current functionalities. However, depending on the configuration of your custom variables, you may not be able to take advantage of this new data model: a clean data model and the power of multi-site queries.
By standardizing these configurations of all your custom variables on all your sites, you will be able to better manage your custom properties. The cleaner your data model, the easier your analyses will be – with no duplication of available properties.
❓ I can't find some indicators in the data model management interface. Why?
In order to provide you with the cleanest possible data model, we list only the custom variables that have generated traffic since January 1, 2018.
Existing data model
The current data model, which feeds Explorer, Dashboards, Reports, Data Query, Reporting API and Data Flow, is site-specific.
There is a shared structure for all your sites, but many properties are specific to each site and in particular:
- custom site variables,
- custom page variables,
- custom order variables.
These custom variables can only be used on page loading events. This data model is therefore suitable for site-to-site analysis but does not allow multi-site analysis on all properties.
Future data model
In a few months, we will be introducing a shared data model for all your sites and applications within your organization to make it easier for you to manage and use the data.
Standard and custom properties
Each event can be assigned all the properties of your data model. The term "properties" will replace the term "dimensions".
The new data model should be seen as a matrix, where the site is only one of many properties:
- in blue: the specific properties of the event, namely the timestamp and its label,
- in grey: the properties of the standard AT Internet data model (OS, Geolocation, page tree, Sales Insights product properties, etc.)
- in yellow: the additional properties you can create. These properties can be used on all types of events, standard or customized.
Moving to the new data model
As you can see, to benefit from this new data model, you will need to transform your current data model, based on custom variables, to adapt it to this new matrix.
The concept is simple: we will retrieve the labels of all your custom variables (sites, pages and commands, as well as your custom global variables) and create a new column for each property. If the same label is used on several sites, or in different types of indicators (sites, pages and orders), they will be shared in a single customised property:
Required action: Standardize custom variable labels
To see your future data models,
If you are satisfied with the properties listed, you are not expected to take any further action, and your future data model is ready!
However, if you notice opportunities for optimization (several properties with different labels that could be shared, for example), read the rest of this article.
⚠️ Priority 1: Standardise customized indicator labels ⚠️
Thanks to your configuration interface, you can harmonise the labels of your custom variables, en masse, on all your sites. ⚠️ This action will immediately change the configuration of your indicators on all your sites.
Caution: if you use the Reporting API, Data Flow API, Scheduled Data Query or Data Flow exports, please note that the labels of the custom variables columns will also be updated accordingly.
Also please note that the rank of the indicators (x1, f8, o5, etc.) has no impact on mutualisation. For example, if you created the Client Type site variable with rank x1 on site A and rank x4 on site B, the new data model will create a single Client Type property.
Variable Label Optimisation Use Case
In this example, the custom variable has been named Product Ref (with "product_ref" as property key) on some pages and ProductRef (with "productref" as property key) on some others.
This setting will create 2 separate columns in your future data model, which is not optimised. We recommend renaming one of them to mutualise as much as possible.
Variable Type Optimisation Use Case
In this case, the Ad id custom variable has been created as a String, whereas it's been declared as an Integer in some others. This format discrepancy generates 2 separate columns in your future data model: "ad_id" and "ad_id_1".
For this kind of optimisation, please get in touch with our Support Centre (firstname.lastname@example.org), your Customer Success Manager or your dedicated consultant as the configuration interface won't allow this format update.
Recommended action: Standardize the values of these variables and your configuration
You can go a step further in the standardisation of your data model. The values of your properties can also be optimized to make your group analyses even more valuable.
To take the example of your identified visitor categories with the following configuration:
|Site 1||Site 2|
|1||Gold Visitors||1||Gold Visitor|
|2||Premium Visitors||2||Premium Visitor|
|3||Platinum Visitors||3||Platinum Visitor|
You can see that on site 2, Visitor is in the singular.
As all the labels are different, a multi-site analysis of your identified visitor categories will return 6 distinct lines:
- Gold Visitors
- Premium Visitors
- Platinum Visitors
- Gold Visitor
- Premium Visitor
- Platinum Visitor
By changing Visitor to Plural to harmonize (Visitors), a multi-site analysis of your identified visitor categories will return only 3 separate lines:
- Gold Visitors
- Premium Visitors
- Platinum Visitors
As with the point mentioned above, the value ID has no impact on the sharing of values. If Gold Customers has ID 1 on site A and ID 4 on site B, you will get a single Gold Customers line in your multi-site analyses.
⚠️ However, it is important to be case sensitive, otherwise the Gold Visitors and gold visitors values will be considered as two separate values, and will be placed on two separate lines.
Here is the list of configurations where optimizing these values on all your sites can be useful:
Configuration of the correspondence tables of the variables of type ID
- Custom organic sources
- Custom marketing sources
- Identified Visitors Categories
- Custom tree structure
- Navigation aisles
- SalesTracker payment and delivery methods
- SalesTracker products
- Monitored IP ranges