The raw clickstream data form the data set that is used by Adobe Analytics. As a service, Adobe can deliver this raw data to the customer in a batched manner on a recurring daily or hourly delivery schedule. This service is called "Analytics Data Feed."
More on what a Data Feed is can be reviewed in the Adobe Analytics support docs here. To get a primer on the Adobe Data Feed interface, check out the Adobe Data Feed video.
Let's get started setting up a data pipeline for Adobe data feeds to your target data lake or warehouse.
Step 1: Logging Into Adobe Data Feeds Configuration
Log into Adobe. Once you are in Adobe Analytics, select the "Data Feed" under Admin:
Next, you will have an option to "add" a new Data Feed:
You will see a screen that allows you to configure your feed. It will look something like this:
Great, you are now ready to se tup a data feed!
Step 2: Configure Your Data Feed
The first part you will need to se tup is called "Feed Information":
Next, you will see options for your "Destination". This is the location that Adobe will be pushing these feeds:
IMPORTANT: The path contains a unique ID that is provided to you by Openbridge. If the path or the ID is not present, your pipeline will not work.
Next, you will want to set different options for the feed. The first is to remove escaped characters:
Next, set for delivery packaging options as shown:
Lastly, you will need to select the columns to be included. We suggest using "All Columns Premium 2018".
IMPORTANT: Once you set a column template, ALL report suites being delivered to the configured destination must follow the same column template. If you select column template for a configured location your data pipeline will fail.
That is it! Your production data pipeline is ready!
NOTE: While not required, we strongly advise reviewing the testing processes detailed in Step 3 and Step 4. Following these steps will ensure smooth configuration and ongoing operations.
Step 3: Testing
This should actually occur prior to Step 2. While you can simply run with a production setup as detailed in Step 2, we suggest testing your configuration in advance of that. This means prior to sending anything into a production data pipeline, we suggest testing your target feeds.
Testing should be done by pushing data to a preconfigured "test" S3 bucket. Nothing sent to the test bucket will be processed. The sole purpose of the test bucket is to allow you to work out the logistics and configuration details in advance.
Once you are comfortable with your test feeds, you can switch them to the production pipeline described above.
Step 4: Ongoing Testing & Audits
Adobe says that their system may occasionally transfer a file more than once. Also, they may have a failure where the data was not delivered.
We suggest that you establish audits of delivered files. Here is a starter header for a template worksheet for tracking your feeds:
FEEDNAME | FEEDID | REPORTSUITE | OWNER | STATUS | DATADATE | RUNDATE
Most are self-explanatory and are available via the Adobe UI. The
DATADATE reflects the dates covered in the delivery. For example, if the feed was covering 30 days, put the covered periods in here. You will want to make sure that all 30 days were delivered and processed.
If there is a delivery failures OR delivery of the same data more than, once you will want to be able to be able to have an audit trail to provide either run the process again OR contact the Adobe support team. We find it is helpful to have an audit and manifest of any gaps/issues once every couple of weeks. If needed, this would be sent to Adobe customer care who can then determine a course of action to remediate any issues that you can not directly resolve. The longer you wait to highlight an issue the longer it will take your internal team or Adobe Customer Care to respond/fix. If the job is old enough, teams may have trouble finding the batch job for a given day.