Originally posted on my employers blog on December 8, 2011: West Interactive Blog
In 1989, Gartner analyst Howard Dresner introduced the term “business intelligence,” which he defined as “concepts and methods to improve business decision-making by using fact-based support systems.” The world has been trying to redefine it ever since.
Let’s say I’m looking to implement a business intelligence solution. I can find lots of definitions for the term business intelligence, or “BI,” depending on where I look. Gartner’s Magic Quadrant for Business Intelligence goes so far as to identify 13 specific capabilities that make up a BI platform, nine of which must be must be delivered for a software vendor to be included in the Magic Quadrant.
Successful notifications campaigns hinge on the ability to deliver relevant, personalized messages to the right customer, at the right time on the right channel. However I try to define it, when it comes to executing on a multichannel notifications strategy, a BI system needs to tell me what has happened and to who, what is happening right now, and what is likely to happen next.
In order to decide the best treatment for a specific customer, I need a historical perspective on what has happened before. I need to understand the key characteristics that apply to each customer, and I need to map that back to what happened on previous interactions. The BI system also needs to be able to tell me which specific customer characteristics are predictors of desired outcomes. This information helps define customer segments and gives me the insight I need to test and apply personalized treatment strategies to specific customer segments.
Next, I need to be able to test the various treatment strategies against subsets of a customer segment to maximize the effectiveness of the strategy. For example, I need to be able to determine which contact channels are most effective to maximize contact rates within a customer segment, and I need to be able test different outbound criteria profiles and message personas to see which ones are most successful at calling customers to action. And, most importantly, I need to be able to see the results and make changes on the fly. If I have to wait too long for the data, then I’ve missed an opportunity both in terms of financial impact and customer satisfaction. Real time is a must.
Knowing what’s going to happen next is the secret sauce that makes a proactive notifications treatment strategy truly sustainable. If I can predict how a customer segment will perform against a specific treatment strategy, I can apply that strategy consistently to achieve optimal results.
A precise definition of BI is not that important to me. Communicating what BI does and how my performance benefits from BI is important. I know I have a winner if I can achieve the following:
- Use individual customer characteristics to define customer segments and match to prior outcomes
- Test various treatment strategies against a customer segment and implement the best one on the fly
- Define a sustainable treatment strategy by predicting how customer segments will perform