Classifications are cool (and sometimes hard to keep up-to-date.)
Classifications (formerly SAINT classifications) are a powerful tool for gaining more insight from your analysis of online behavior. They provide a means to upload metadata that can be connected to the information being gathered by Adobe Analytics. Classifications are used in many ways. Here are a few examples:
- Classifications are often used to break a variable containing a delimited list into smaller pieces. For example, tracking codes often comprise a delimited list with portions of the tracking code string representing marketing channel, partner, name, date. etc. Classifications can be used to group tracking codes by these categories.
- Some retailers populate the Products variable with a SKU number. Using classifications they can upload information from their internal product database to categorize SKUs into brands, sizes, colors, flavors, varieties, categories, etc.
- Some companies populate a prop or eVar with a customer ID and use classifications to upload customer attributes for from their CRM database for use in Analytics.
- Classifications can be used to give friendly names to cryptic codes stored in a prop or eVar. For example, you can use classifications to rename ‘ca’ to ‘Channel Affiliate.’
But there’s a dirty little secret about classifications: They can be very time consuming to manage. As an example, the typical process for updating classifications for tracking codes might look something like this:
- New tracking codes are created and used in online advertising efforts.
- The person managing classifications notices that these new codes are not classified properly. Often this is exhibited in reports by the “None” bucket in a classifications report growing larger and larger over time as more and more unclassified tracking codes come into the system.
- The person managing classifications downloads the existing and unclassified keys via the classifications export tool in Analytics and loads the file into Excel.
- Using Excel she types the correct classifications in each cell for every unclassified key, saves the file, and then uploads it back into Analytics using the import tool.
- She waits for a few hours or a day or two for the classifications to propagate to reports.
- Voila, tracking codes are appropriately classified in Analytics.
- A few days go by. More unclassified tracking codes come into the system.
- Go to step 1. (sigh)
The process is manual, slow and often reactive in nature which means classifications reports are not as useful as they might otherwise be. Sheesh, why can’t we at Adobe make this easier??? Well, we’re working on it. We have several initiatives underway to improve the situation. This post is Part 1 of a multi-part series describing one of those efforts: the classifications rule builder.
Automating classifications with a rule-based approach
As indicated above, classification metadata can come from many places. The new classifications rule builder focuses on use cases where metadata is located within the string values being passed into Adobe Analytics. The tracking code example mentioned above is a prime example of this use case. Other examples include certain product strings and internal search terms. Using rules you can tell Adobe Analytics how to decipher the incoming tracking codes, search terms, etc. and then Analytics will automatically set classifications for you. You no longer have to deal with downloading, filling out, and uploading classification files all the time.
In my next post I will talk through several examples in more detail and show how a rule-based approach will greatly simplify your process, saving you hours of work each.
The classifications rule builder is available NOW
For those of you who attended the Digital Marketing Summit in Salt Lake City or London, the classifications rule builder was demoed during the closing SNEAKS session at those events. I’m happy to announce that the classifications rule builder is available now for use by all Adobe Analytics customers.