Over the last three weeks, we have looked at CJA Golden Nuggets, Tips & Tricks, and Best Practices. Today we will look at integrating CJA with other data and tools that can unlock new insights for your organization. Here we’ll look primarily at bringing data into CJA. We’ll touch only briefly at ways to get data out of CJA.
Overview
One of CJA’s benefits is the ability to analyze online and offline customer engagement holistically. Having a clear understanding of CJA’s datasets, requirements, and how to integrate the data is important to ensure the data is useful in addressing key business questions.
Use cases
There are almost a limitless number of use cases where data can be integrated with CJA. Today we’ll focus on three typical use cases.
Customer service interactions – For many companies, the ability to better understand the full customer journey is essential. Being able to see person-focused web and customer service interaction seamlessly would unlock a great amount of insight. In this use case, we’ll add an interaction level CRM dataset to our web data.
Product metadata – This dataset will provide a variety of datapoints for individual products. It will allow end users to understand larger trends. We’ll add a csv file with multiple attributes for our top products.
Profile data – In our final use case, we’ll bring in person-level profile data that will make available new ways to analyze our web and customer service data.
AEP Sources
Adobe Experience Platform has a number of source and destination connectors available. Sources can simplify the process of bringing data into AEP. Destinations, on the other hand, allow you to send data from AEP downstream.
Dataset types
Currently CJA supports three dataset types. We’ll look at an overview of each and discuss the requirements.
Event
Event datasets provide details on events that occur at a certain time. Clickstream data is the prime example of an event dataset. Each page visited, perhaps each button clicked, and certainly each item added to cart is represented, often with an enormous amount of metadata. Our customer service interaction dataset mentioned above will be an event dataset. Other examples could include in-store sales or travel data such as flights and visits to the airline VIP lounge.
Event datasets require a timestamp and a personID. Until recently, the personID had to be consistent with the web data. CJA recently introduced graph-based stitching that can build relationships between multiple IDs . Data distiller can also be used for custom stitching options. (Full details of stitching are outside the scope today, but details can be found in Adobe’s documentation.) Event datasets can contain many rows for each interaction, such as web data. Or the entire interaction may be reported in a single row, such as our example CRM customer service data.
Lookup
Lookup data requires a matching key to event data. For those who have worked in traditional Analytics, lookup datasets are like classification files. A prime example is product metadata using SKU as the matching key. When a visitor interacts with a product online or via customer service, event data records the SKU. The metadata dataset will provide a variety of attributes for each SKU such as description, category, subcategory, and vendor.
Profile
Profile datasets are based on the personID setup for event datasets. These are like traditional Analytics customer attributes. When the profile dataset contains a personID matching an event, the appropriate attributes will be reported. Our profile dataset contains attributes for each of our known customers.
Important notes
In CJA reporting, attributes from lookup and profile datasets will always return the last updated data that matches the event data . The lookup and profile attributes are not based on when an event occurred. For example, let’s say the profile dataset includes customer_tier. If a customer moved from silver tier to gold tier on May 1, gold tier will now be reported for all events related to this customer – both before and after May 1.
Data in Workspace is based on the reporting window selected. These dates determine which events are included in the report. Only lookup and profile data that matches event fields will be reflected in the reporting. Again, as noted above, the lookup and profile data aren’t date specific. The most recently ingested matching rows will be reported regardless of when the event happened.
When you have more than one event dataset, if the schema paths match, the fields will be combined in CJA. For example, if revenue from multiple datasets is captured in _experience.revenue, the revenue field will be combined. On the other hand, if one dataset is _experience.revenue and another is _experience.commerce.revenue, those fields will not be combined.
Adding datasets to connection
Now that we’ve had an overview of the three types of datasets in CJA, we’ll look at the process of adding this data. We’ll assume that your web data is already set up.
Once you have each dataset in your AEP sandbox, it’s time to add them to the CJA connection. This process defines the matching key and allows you to select backfill options. Full details can be found in Adobe’s documentation.
Adding to the data view
After your new datasets have been processed, you’ll want to add the desired components to the Data View and validate the data.
As best practice, you should include clear descriptions of your new attributes. As mentioned in previous weeks, clear descriptions allow users to more effectively self-serve. Additionally, users should know how to identify the source of particular attributes. For example, perhaps the product metadata contains a vendor field. An effective description for the vendor field might be something like: “Vendor for the product related to the event. Based on event SKU.”
Since SKU is the matching key in this example, you’ll have a SKU from event datasets and the lookup data. Typically, I’d suggest hiding or removing the SKU field from the lookup data. While it may be useful for validation and troubleshooting, duplicative fields can be confusing to end users. Retaining the event SKU ensures that the value is available even when it doesn’t exist in the metadata.
Coming soon: summary datasets
Adobe has announced that summary datasets will be supported in the future. Those familiar with Adobe Analytics may know data sources can be valuable to report on various data such as email data by campaign, offline sales when event data isn’t available, or any other summary data. Details haven’t yet been released, but I’d expect this to work like it does in traditional analytics.
Using email data, for example, users may be able to view the number of emails sent and opened by day and campaign. This data can be a great supplement to web traffic data for those same days and campaign.
Data budgets
In CJA, it appears each row uploaded for a summary dataset will count toward the allocated data budget. Rows for profile and lookup datasets also count toward the allocation. In traditional Analytics, there’s no extra usage incurred by data sources or classifications. The number of customer attributes is based upon the Analytics package an organization has purchased.
For heavy users of these summary data sources, the rows could quickly reduce the amount of data available for event reporting or cause an organization to incur overage charges. Uploading profile datasets for large amounts of customers can quickly add up. The same is true for daily summary datasets when broken out by tracking code. Since tracking codes can have many permutations, these summary datasets can easily add up to tens of millions of rows a year.
As is the case in traditional Analytics, wouldn’t it be nice if Adobe included the ability to leverage these enrichment datasets in CJA without needing to worry about overage charges?
Exporting data from CJA
Finally, we’ll look at several ways to export data from CJA so it can be used in other tools. There may be limitations based on which tier of CJA your organization is using. Refer to the linked documentation for more details.
CJA offers several new and improved approaches.
- Standard Workspace exports – you can easily export and schedule PDF and CSV files of Workspace projects. There are also several options to right-click and copy or download portions of projects.
- Full table exports – allow exporting tables with millions of rows, up to 5 metrics, and up to 5 breakouts. These exports are sent to cloud locations or a temporary Adobe Landing Zone location where they can be retrieved or used with other tools.
- BI Extension – a PostgreSQL interface that allows data to be used in tools such as Tableau and Power BI.
- Report Builder – an updated Excel plugin that allows you to bring data from CJA into Excel. Once the data is in Excel, you can you all the typical Excel functions.
Adobe also discusses various data export use cases and provides several suggestions.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.