Understanding customers is a preoccupation for many BI experts. Slicing and dicing information to find hidden trends and make better decisions is a big part of what BI has to offer, but how do you apply this knowledge at the point of interaction? Traditional BI offers some challenges in this regard. First, BI is primarily about the analysis of historical information. It may be as fresh as yesterday, but it rarely is focused on the present moment. For analysis, BI must consume a fair amount of historical information in order to prove or disprove a hypothesis. Traditional BI is not good at dealing with the here and now and making decisions that deal with one specific customer.
Setting up the proper environment, figuring out the right offer and making the most effective pitch is the front line of the consumer products and services world. All the work that goes into understanding your customer is focused on making sure things go smoothly at the point of interaction. This is where you make an impression and hopefully a sale, and it is the convergence of operational and analytical processes for your business. To make my point, allow me to share a personal story.
As a frequent business traveler, I have reached a high-level customer status with a major U.S. airline. One winter, I was traveling on a flight from Baltimore to San Jose, with
a stopover in Denver. Upon arriving in Denver, I found that nearly every outbound flight was either delayed or canceled due to a severe winter storm - including my connection to San Jose. Then, I noticed on the departure screen that another flight to San Jose had been delayed but was scheduled to depart in less than an hour. I headed quickly toward the new departure gate and found myself sixth in line, watching fellow travelers approach the ticket agent with their requests to board the soon-to-be departing flight. One by one, they were politely but firmly denied. Then it was my turn. I sheepishly approached the agent to make my request, fully expecting to hear another firm denial. Yet, this time, she looked at my current flight boarding pass, noted my airline high-level status and promptly reissued a new ticket for the departing flight. She also managed to honor my upgrade request and have my luggage pulled and diverted so it would arrive with me. When I regained my composure I asked her, "What just happened here?" She smiled and said, "Exceptions can be made, in particular for our high-status customers."
Here lies the point of interaction where a customer and the business meet. In this case, the agent knew what to do and, more importantly, had the most up-to-date information from which to act. Arguably, airlines are not always prime examples of customer service, but at this particular point of interaction, the company made a positive impression on me, and as a result, I have maintained my loyalty to them ever since.
The Role of Analysis
From a business perspective, a considerable amount of analysis and work go into profiling and understanding the buying behavior of customers, and where the point of interaction is dynamic, particularly for online transactions, analytics play a critical role. Offers in the form of banners, suggestions for future purchases or even complementary product offerings are all geared toward a recipient based on his/her past buying behavior and other demographic information.
Figuring out the appropriate approach at the point of interaction requires information about your customer. In the past, a successful car salesman might size up a customer based on appearance and pick his approach based on past experience. In the modern world, we've supplemented experience by sizing up customers electronically and introducing some straightforward clues that help guide the interactions between sales representatives and customers. If a customer belongs to your discount club or holds a status card, then the representative does A; if not, the representative does B. But a club membership by itself doesn't always provide enough information about the customer to enable the representative to make a consistent value-based decision.
This is where analysis comes in. When you have a fairly sophisticated point-of-sales system or an online presence, you can leverage immediate outcomes. With a BI tool, an individual customer's behavior can be easily extracted and dissected. Then, an experienced analyst can find trends that will help predict future behavior or additional interests. Naturally, this leads to decisions on what types of offers to make, approaches to making the offers and other ways to improve the customer's experience.
The Problems with BI at the Point of Interaction
To determine whether customers who buy product A also tend to buy product B, you'll need to identify a historical trend that demonstrates a correlation. Traditional BI tools are very effective at finding and analyzing historical trends to help improve a decision-making process, such as decisions around revenue trends or the popularity of a particular brand. But traditional BI does not deal well with the present, nor with decisions that relate to a specific customer. Whether you use dashboards or OLAP or ROLAP or MOLAP, these tools require a fair amount of business acumen, if not technical capability. This often results in more time spent hunting for information than analyzing it. At the point of interaction, the customer representative should be prompted to act based on a simple exterior clue. In my original example about the airline, the ticket agent's clue was my status printed on the boarding pass. For others, it could be a notation on the checkout screen that becomes the basis for an action. In any case, it needs to be simple and easily understood. However, just because the exterior clue has to be simple does not mean that the underlying analysis is also simple. BI provides a wide variety of sophisticated tools for analyzing data and predicting future trends. Wouldn't it be great if you could programmatically apply those same tools to the present point of interaction?
Building a BI Environment for the Point of Interaction
Let's look at the example of a bank. In today's highly competitive financial industry, customer retention/attraction programs are key reasons why banks are interested in predicting which customers might switch over to the competition. Analysis of a customer's transaction history is one such method, and this is where operational BI comes in. Because a historical record is vital to predicting behavior, a traditional data warehouse or a data mart can be utilized. The traditional BI analyst would examine the aggregate of the customer's data, searching for behavior that signals someone is about to close an account, such as:
- Stopping direct deposits,
- Transferring large, steady amounts of cash to another (outside) account or
- Recent closure of another account, such as a credit card.
Reviewing this historical information, the BI analyst can then prove or disprove his/her hypotheses to create a model for predicting behavior. One question that may crop up: If I have the model, why not just tag the customer record; why do I need to make the analysis operational? The answer is that the current transaction may be the most critical piece of information for your analysis, and you want to make the decision at a time when you have the best opportunity to influence the outcome. While historical information is useful in certain cases, it is not as relevant as information that tells you what is happening in the present.
Delivery of Information
The final step, delivery of this information, is critical. When training an employee - or the system - to behave at the point of interaction, the process will be more streamlined with fewer variables. A simple red box or highlighting of a name can be enough of a clue to the service representative, allowing him/her to take a specific, predetermined action. One requirement is that your operational BI system should be easily incorporated with your point of interaction system. Therefore, using a system that is based on an SOA or clearly defined user interface is important. Another key element is the ability to translate analysis into operations. If the analysis done to generate a predictive model is not easily carried into the operational system, you will likely lose the flexibility to make changes to the system as new information presents itself.
Unlike traditional BI, an operational BI system should be focused on influencing the interaction with your customer to provide benefit to both the customer and your business. Traditional BI, while often seen as a tool with a very fuzzy ROI, is nonetheless necessary for conducting business. Operational BI, on the other hand, provides a much clearer benefit because it directly addresses your business.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access