Identifying the right data to support a journey analytics use case is the most critical component to the success of the entire program. There are many moving parts in a journey analytics program, such as: executive sponsorship, the team, the competency and skill sets of the team, the right goals and objectives, but none is more important than finding the right data.
The data used in journey analysis must contain enough information to illuminate the story about how the customer navigated through the various channels. Business processes can change a customer record or route them in many different directions. These are all clues to the puzzle that enables a customer to fulfill the purpose of their journey.
There are four major classes that comprise the foundation of data used in journey analytics initiatives. Data from each class convey different parts of the story. Here's a description of the four data classes:
- Interaction Data – Data collected by the systems, or channels, in which a customer directly interfaces. (e.g. A website interaction by the customer to pay a bill.)
- Transaction Data – The recorded details and outcome of a customer transaction with a company. (e.g. A credit card decline and the details about why it declined.)
- Business Data – business data that can be correlated to interaction or transaction data. (e.g. Satisfaction, complaint, or dispute records.)
- Segmentation Data – Attributes about the customer. (e.g. Demographic or Subscribed Product information.)
Journey Data's Many Forms
From raw system logs to data warehouse table extracts. There are some simple rules on where to find the best data source per data classification:
- Raw System Logs are the most robust interaction data source to tell an exhaustive story of how a customer navigated the various channels during the interaction. These logs were designed for verbose debugging and typically have the deepest level of detail about a customer interaction. These can be considered "messy" data.
- Data Warehouse & Database Table Extracts will commonly house transaction, business and segmentation data. These data types typically have business logic applied upstream prior to the data being written to the database and is often summarized. These data sources are generally "cleaner" and better organized.
Even with these fairly simple rules on where the necessary data might live, it’s often a challenge to find the right person in an organization that can locate the right data for an initiative. Especially when you're looking to collect data from several channels for a large journey analytics initiative. This speaks to data, like organizations, being organized in silos by business unit, as well as the complexity of data.
For raw system log data, in our experience, the most effective way to acquire the right data is by finding the data producer. This is the Subject Matter Expert (SME) who knows the system that generates the data in its native environment. Getting to data in its native form ensures it has not been cleansed, typical of data warehousing processes, and that it contains all of the elements required to tell the unaltered story of the journey. For example, an IVR developer is best suited to provide IVR data, while web developers or web analytics teams are best suited to locate web data.
On the flip side, the data warehouse and database table extracts traditionally come from data consumers. Data consumers are those who use data that has already been cleansed. Data cleansing, typically, has little impact on the power of the data when combined with other sources to form journeys. These extract sources are normally much lighter in size (structured, delimited text, as opposed to verbose system logs) and contain more focus around each category. For example, transaction data will contain: transaction type, amount, account, date, time and other transaction attributes.
Warehouse extractions represent the most common data source we see, while raw system logs will generally include more comprehensive detail.
This is customarily the hardest part of the data discovery process – finding the right data sources and their experts. Once the right data is found, the foundation is built for a successful journey analytics program.
In the next blog, we’ll explore some of the best data sources for each category.