Decide: Making a choice in the business moment
In this, the fourth installment of our five part series focused on real-time actionable insight, we explore how to optimize decision management by infusing context-aware streaming analytics into all decisions. Using IBM’s four pillar approach (Sense, Orient, Decide and Act) you can be assured that your decision will result in smarter business processes and more intelligent operations.
Humans make decisions every day. Some decisions are highly automated: for example, ducking for cover when the neighbor kids launch a fly ball into your yard or sweating as you power through that last quarter mile on the treadmill. Other decisions are more deliberate and require a bit of brain power: Paper or plastic? Skim, soy, 2 percent or whole milk with your latte? Still other decisions are highly complex and challenging, such as how to manage your financial portfolio or how to help your teenager make smart life choices. These more complex decisions exercise your brain power, build on context and depend on sophisticated inputs.
Your business also has tough decisions to make which are fundamental to keeping enterprise operations up and running. For example, where on your e-commerce site homepage is it best to display customer login inputs? Other decisions, such as which companies to buy or which business units to divest require analytical rigor and significant time and attention.
It is interesting to note that humans feel less stressed and make better decisions when there are fewer choices. Decision fatigue is a threat to senior leaders and refers to the deteriorating quality of decisions after a long session of decision making. It is believed to be a root cause of irrational trade-offs. For instance, judges in court have been shown to make less favorable decisions later in the day than early in the day.
Organizations are looking to technology to cut down on decision fatigue through automation of operations and business processes. A human should be involved only as required and, when a human decision maker is involved, the choices should be clear and refined, based on sophisticated context-aware, real-time analytics of streaming data.
Decision makers should also have the tools and data to explore the impact of their decisions and perform real-time analysis on “what-if” scenarios.
Big data can help you decide
From my experience attending and speaking at many big data, integration and governance forums, I have found that forum attendees are consistently turning to big data solutions to make better decisions around customer management. Some of the questions I often hear are:
- Which clients to retain?
- Where to find new clients?
- What offers to send to whom, and when?
The good news is that IBM’s four pillar approach—Sense, Orient, Decide and Act—is helping organizations fuse diverse data sources and analytics to bring context and insight, reducing decision overload and automating the best possible decisions.
For example, IBM is helping retail clients capture a new market opportunity: the “in-range” customer. Imagine a customer riding the subway or bus home from work and wondering what to prepare for dinner when she receives a promotion on her phone from the neighborhood grocer offering a great deal on a delicious, ready to serve chicken dinner. Using the IBM Big Data & Analytics portfolio (which employs real-time geospatial data, predictive models and business rules) the grocer now has access to real-time actionable insight, enabling more targeted customer offers, sending the right promotion at the right time to sell more chicken dinners.
How could real-time analysis of streaming data sources enhance traditional business rules, which aren’t able to handle new streaming data types such as geo-spatial position? Think about log files from your data center, machine generated outputs from sensors on manufacturing equipment or data streaming off the smart grid—these are all sources for customer data that could change the way you do business.
Bringing together context, analytics and decision management
Real-time actionable insight helps organizations optimize decisions and implement repeatable business outcomes across all processes, applications and interactions in the business moment. It does this by:
- Sensing every data point and event to capture what is happening
- Putting data and events into context to understand and evaluate how everything relates
- Applying analytics and business rules to gain best the possible insight to decide what is best
- Putting that decision into action where it is needed the most—in processes, applications and interactions
Real-time actionable insight turns insight to action. How? By bringing together context, stream computing, predictive analytics and decision management.
I think its worth focusing on a few key terms to really highlight the depth of this vision:
- Real-time actions or decisions could mean milliseconds, or even microseconds; real-time could also mean completing an action before a deadline or event (right time).
- Action is responding in a meaningful way (automated or manual) or making an informed decision—aligning responses by people or systems to desired outcomes based on plans, activity and resources. As any of these elements change in real time, so too must the prescribed action to be taken by people or systems to continue ensuring optimal outcomes.
- Predictive is data-driven optimized decision-making using past and present data to forecast future outcomes and behaviors, as well as understand what drives them.
- Context is the ability to better understand something by taking into account the things around it—across ALL entities—and understanding how an observation (data) relates to your business.
How the pillars work together to fight fraud
Let’s explore an example of how “sense, orient, decide and act” work together by discussing fraud detection and prevention. The rising cost and complexity of fraud and money laundering result in brand damage and financial ruin.
- Mortgage Industry: The cost of fraud is estimated by the FBI at more than $1 billion nationwide
- Healthcare: The National Health Care Anti-Fraud Association estimates the healthcare industry’s cost of fraud at more than $60 billion
- Corporate Sector: The Association of Certified Fraud Examiners estimates the cost of fraud in U.S. organizations at 7 percent of annual revenues, or $994 billion
- Insurance Industry: The National Insurance Crime Bureau estimates the cost of fraud at a total of $20 billion a year
The solution comes in the form of multiple automated actions available to keep fraud under control, in real time:
- Automatic creation of fraud cases for security analysts to investigate
- Alerts at point of sale to prompt deeper investigation including requests for additional proof of identity (photo ID) or a waiting period before transaction completion
- Real-time alerts and updates to existing fraud cases
- Automatically build a list of actions for security analyst so they know exactly where to focus, for example: placing a person on black list or watch list
- Automatically stop transactions in the case of a compliance violation
Tighter integration between business rules and real-time analytics represents exciting opportunities for businesses. The result of this marriage is a growing set of dynamic business rules built on a rich and sophisticated framework of analytics such as machine learning, cognitive computing and text analytics.
This integration will bring about the next generation of decision management. An example might be better taxi management through automated fare adjustment and optimization of routes. The goal is to keep the business focused on opportunities while driving down risk and minimizing time spent chasing the wrong opportunities.
Next week we will close out our series with a discussion of “Act,” the fourth pillar. Join us then, and, in the meantime, engage with us on Twitter using #IBMchickendinner and enjoy this reading:
- Streaming Analytics in Industry
- Convert Real-Time Insight into Action
- The Nature of Analytics Video Series