One of the topics that are of increasing interest to the web analytics community has been the analysis of multi channel data and how organizations can most effectively leverage this process for optimal value. This month, we will explore the business case for investing in multi channel data analysis, the highest value channels of data to integrate together and the best practices for approaching this type of analysis.
The reason why the analysis of multi channel data is so critical is quite simple. If you are not able to effectively analyze your customer interactions from all points of contact then it is highly likely that you will make incorrect decisions based on incomplete customer information. Take for example the table below which illustrates a very simple problem that is extremely common:
|Customer X, Campaign 1|
|Campaign Type||Paid Search|
|Search Term||"Khaki Pants"|
|Revenue generated from online purchase||$40|
|Revenue generated from in-store purchase||$0|
|Total revenue generated as a result of customer X viewing campaign 1||$40|
|Customer Y, Campaign 2|
|Campaign Type||Banner Ad|
|Campaign ID||"Khaki Pants"|
|Revenue generated from online purchase||$20|
|Revenue generated from in-store purchase||$100|
|Total revenue generated as a result of customer Y viewing campaign 2||$120|
In the example above, if you are only able to evaluate the online purchase data, then you would incorrectly conclude that Customer X in Campaign 1 produces twice as much revenue as customer Y in Campaign 2 ($40 online in Campaign 1 versus $20 online in campaign 2). In reality, after considering all of the customer interactions within all of the available channels, Customer Y in Campaign 2 actually produces three times more total revenue across all channels than Customer X ($120 versus $40). What this simple example illustrates is that unless you are tracking every channel of interaction with your customers you are running the risk of making poor decisions based on incomplete information.
High Value Data Sources
A key starting point in the multi channel analysis process is identifying the sources of data that should be integrated first. In an effort to include all of the relevant channels of data into your multi channel analysis, it is important to start with the most important data sources for customer information. Some of the most common high value channels of data include the following:
Multi-Channel Data Analysis Best Practices
In an ideal world there would be a common element within each different channel that would ease the process of correlating data sources from various channels. In some cases organizations do leverage this method by enticing a large population of their customers to use discount or loyalty cards (frequent flyer cards, preferred customer cards, etc.), which have the ability to ease the process of tracking customers actions across multiple channels.
However if your organization is not currently able to leverage this type of program to merge your data across channels this does not mean you cannot experience real value from multi channel data analysis. It is critically important to have some method for correlating the data between channels. In some cases, simply being able to segment and correlate channel data by geographic region or by linking call center sales to a specific web channel marketing campaign (via a unique 800 number) is all that is required to derive significant value from the analysis of multi channel data.
Another common misconception is that in order to gain value from multi channel data, all of your data must be combined within a single huge data warehouse for the analysis to take place. In reality, 80% of the value from multi channel data is often available from only 20% of the effort by leveraging 2 separate data warehouses. One specializing in offline data (often already in existence) and one that is newly built, based on the web channel and related data. This approach allows each system to be configured to exchange high value summary data that can easily be integrated into the other warehouse.
The net result is having specialized environments where each system can be leveraged to answer questions based on which source of data needs to be more heavily leveraged. If the question you are trying to answer pertains mainly to in store behavior, for example, "What was the in store average order size for customers who participated in online program A?" This question would be most efficiently answered within the offline warehouse by importing the needed segment information from the online warehouse. On the other hand, if your question was "What are the most popular products purchased online by customers who shop via catalog, in store, and online?" Then this analysis would most efficiently be done using the warehouse focused on online behavior.
At the end of the day those businesses that interact with their customers via multiple channels will need to leverage multi channel data analysis to truly have a 360 degree view of their customer and to gain a competitive advantage in the marketplace.
Bill Bruno is the CEO - North America, Ebiquity.