Analytics in a Multichannel World
CRM is increasingly being practiced in a multichannel world. Customers might use a branch or store and still visit the Web site. They may use kiosks or ATMs and still call the contact center. They may opt out of email but opt in for text messaging. To manage the customer relationship we must be able to ensure an appropriate customer experience across all these channels. Two characteristics of a multichannel world make this effort more complex:
- Many of these channels are automated -- there is no one there to build a relationship; and
- deciding how to use each channel with different customers at different times can be difficult.
Effective use of automated channels to build customer relationships requires that those channels do more than merely display information -- they must take appropriate actions. They must respond effectively to customer requests and actions and they must be proactive in making offers or suggesting alternatives. Just as a good contact center representative finds the right next step for a customer, so must these automated channels. If these channels are to take actions then they must be made capable of good decisions about what actions are appropriate, allowed, and likely to be effective.
Focus on the decisions that must be made in these channels, and it becomes increasingly apparent that making decisions based only on our existing customer data is limiting. For instance, knowing the number of minutes a customer used her cell phone last month might help us decide about an appropriate upsell offer -- but we would likely make a better decision if we could see that next month would continue her recent upward trend. Knowing how far behind she is on her payment plan might let us select the right collections approach -- but knowing how likely she is to self-correct in the future would be more useful. Knowing her contract is coming up for renewal might be enough to decide on the right retention offer -- but knowing that customers similar to her respond better to free gifts than to extra minutes would improve our chances of making the right decision.
Enriching the set of customer information to enable better decisions requires analytics. Analytics can segment customers into like-minded and like-acting groups to improve our focus. Analytics can also turn uncertainty about the future -- especially future risk and future value -- into probabilities. This allows us to assess how likely this customer is to be risky or valuable in the future. In these ways analytics can make for more-precise decisions and more-effective actions.
Because we are talking about automated channels, such analytics cannot rely on visualization, Excel, dashboards, or other staples of business intelligence. To push analytics into our automated channels we need to deliver this analytic insight in ways that can be consumed by our information systems. We must use data mining and predictive analytic techniques that have executable outcomes such as decision trees, scorecards, and regression formulae. We must also know which decisions we are hoping to improve with these analytics and we must be managing these decisions in a way that allows us to inject our analytic insight. To ensure consistency across channels, we should not embed these decisions in specific channels but manage them as cross-channel, enterprisewide decisions. Only then can we take full advantage of our analytics.
Analytics have a second role in a multichannel world: Multiple channels increase complexity and better analytics are required to maximize our effectiveness. While managing lots of channels means integrating them and being consistent across them, it is also essential that we drive customers to the most-effective channel for each activity. Customers have preferences about channels but the organization has a need for effective channel use. Using analytics to see which channels work for which customers (or customer segments) for which action maximizes our effectiveness. In a collections environment, for instance, a customer may prefer email but our analytics may tell us that email is ineffective for that customer segment. A particular channel -- say, SMS -- might be very cheap, but analytics may show that the cost/benefit of a more-expensive channel can be justified for a particular group of customers.
Because we need to use multiple channels to manage our customer relationships, those channels must take appropriate actions and be managed for maximum effectiveness. Analytics that enhance the data we have to work with and help us see into the future can make all the difference.
About the Author
James Taylor (james@DecisionManagementSolutions.com) is a recognized expert and independent consultant in decision management. He published the book Smart (Enough) Systems with Neil Raden and is a regular blogger and writer. For more information go to DecisionManagementSolutions.com/CRM
Please note that the Viewpoints listed in CRM magazine and appearing on destinationCRM.com represent the perspective of the authors, and not necessarily those of the magazine or its editors. You may leave a public comment regarding this article by clicking on "Comments" at the top.
To contact the editors, please email editor@destinationCRM.com.
To subscribe to CRM magazine, please visit http://www.destinationCRM.com/subscribe/.
If you would like to submit a Viewpoint for consideration on a topic related to customer relationship management, please email viewpoints@destinationCRM.com.
For the rest of the July 2009 issue of CRM magazine please click here.
Putting CRM to Work
Actions, not information, are the true mark of intelligence.
Neither your best guess nor your gut instinct is good enough anymore. Fortunately, with enterprise decision management, the technology and methodology now exist to help you reach a better conclusion.