Today's business trends are driving organizations to build systems smart enough to cope with the demands of a more complex world.  As a result, organizations must automate and improve far more of the decisions underpinning day-to-day business operations.  They must treat these decisions as a corporate asset in the same way they treat their data as one -- perhaps even more so.  These decisions are too numerous and cumulatively too important to be handled in an ad hoc manner.  Decisions (or at least the definition of a good decision) change rapidly and influence your organization's behavior.  A new systematic approach called enterprise decision management is needed.

Introducing Enterprise Decision Management

Enterprise decision management (EDM) is a systematic approach to automating and improving operational business decisions.  It aims to increase the precision, consistency, and agility of these decisions and reduce the time to decide and the cost of the decision.

To make your systems smart enough, your core problem is knowing what's the right decision to make and how to make it when required.  These decisions must be precise (accurate and informed by past experience and actual data), consistent (across staff and channels), and yet agile (easy to change when necessary).  Making your systems smart enough shifts the focus away from manual and hidden decisions and toward decisions embedded in systems, as shown in Figure 1.  Before adopting enterprise decision management and using it to deliver smart enough systems, most operational decisions are made by people, with only a few embedded decisions.  Many operational decisions aren't made at all but remain hidden.  After adopting enterprise decision management, the balance shifts, with more operational decisions identified and embedded in systems, many manual decisions automated, and the opportunity for people to make new decisions of more value to the business because they are freed from operational decision making.

Figure 1.  The change in operational decision making when enterprise decision management is adopted.

The EDM Process

Enterprise decision management involves a straightforward approach, at least in principle.  The basic process is as follows:

1. Remove decisions from applications.  Decisions can't be managed effectively, or even automated effectively, while they're embedded in other applications.  You must remove decision making from your applications.
   
   
2. Analyze decisions' potential for improvement.  Realistically, not all decisions lend themselves to the same degree of automation and management.  In a later instalment on "The ROI for Enterprise Decision Management," benefits and return on investment (ROI) will be discussed.  Decisions offering the highest potential return should be addressed first.
   
   
3. Automate key operational decisions.  The key operational decisions, separated into what are called 'decision services', can then be automated.  Typically, this process involves a set of business rules developed from knowledgeable staff, regulations, policies, and perhaps legacy code.
   
   
4. Apply predictive and decision analytics.  Applying analytics brings insight derived from historical data to bear on your decisions.  This step might take place in parallel with the previous one or come later, as a way to enhance decision making.
   
   
5. Give business users control.  At some point in the process, you will hand over control of all or some of the decision logic to business users in your organization.  This change of focus is essential for ensuring maximum agility and collaboration.
   
   
6. Keep it simple as 'intelligence' increases.  The more complex decision making becomes, the more important it is that the business can control the logic without too much of the complexity showing.
   
   
7. Focus on production performance requirements.  The decision process is running in production, so you must focus on performance and reliability the way you would for any operational system.
   
   
8. Manage change and evolve.  No decision is ever perfectly automated, or at least not for long.  You must develop a mind-set of constant testing and refining the way you make these decisions.

The EDM Definition

You can consider the three parts of the phrase "enterprise decision management" separately to understand the concept better:

• Enterprise.  You should treat decisions as an enterprise asset.  You should make even the most high-volume, operational decisions as though they are enterprise decisions.  Your customers and other associates behave as though this approach is the case.  When you decide to contact them or treat them in a certain way, they assume that decision was deliberate on your part.  In fact, most organizations not using enterprise decision management make many decisions incidentally, without thought, or implicitly.
  

'Enterprise' is also the 'E' word, possibly the most overused term in the software development world.  Every product and every approach seem to be 'enterprise' class.  Despite this, the term is part of the phrase because enterprise decision management is most effective when you can cross information 'silos' and use all your information to inform decisions and share common decisions throughout your application portfolio.  Most organizations operate in silos, with different parts of the organization 'owning' different data and making decisions separately.  Enterprise decision management involves focusing on decision making as an enterprise asset, not as a departmental one or an individual one.  Customers expect the decisions an organization makes to be deliberate, not accidental.  The 'E' in enterprise decision management is about making sure this approach is followed.

Note.  In general, you'll be most successful if you can make use of a common decision-making infrastructure across silos and systems, but enterprise decision management doesn't have to touch every part of your organization. 

You can apply EDM techniques to a single decision within a single silo and get a great deal of value.  If you can apply EDM across multiple silos there is a multiplying effect that will likely be a large part of the value and benefit you get from EDM in the long term.  So think 'enterprise' even if you start with a single project.

• Decision.  A decision is a determination arrived at after consideration, a solution that ends uncertainty or dispute about something.  It's the act of selecting a course of action.  Therefore, decisions
  • Select from alternatives, typically to find the one most profitable or appropriate for the organization and/or associate.
  • Consider various facts or pieces of information about the situation and participants.
  • Result in an action being taken, not just knowledge being added to what's known.

Organizations make decisions at every level, from strategic decisions the CEO makes to operational decisions customer service representatives (CSRs) or store workers make.  Other decisions must be completely automated (for instance, in customer self-service applications), or the people making decisions need a system that helps them do so (such as a call center application supporting call center CSRs).  Typically, enterprise decision management focuses on operational decisions.  As an approach, it's intensely decision centered rather than process or system centered.

• Management.  It's called enterprise decision management, not enterprise decision automation or enterprise decision improvement, for a reason.  Management is about conducting or supervising something and the judicious use of a means to accomplish an end.  Management in an EDM context means treating decision making as a business issue and a business asset, focusing on accomplishing an end (taking an action) and on supervision to ensure that improvement and optimization of decisions are ongoing and proactive.  You can't just automate decisions and hope for the best; you must manage them over time. 

[Sidebar-1.  Large European Bank:  Risk Management and Compliance]

The following list explains some critical phrases in the formal definition of enterprise decision management you saw at the beginning of this article:

Enterprise decision management is a systematic approach to automating and improving operational business decisions.  It aims to increase the precision, consistency, and agility of these decisions and reduce the time to decide and the cost of the decision.

• systematic approach -- Enterprise decision management is a planned, methodical approach to automating decisions.  It's not about automating decisions as a side effect of another activity, such as automating a process or implementing an enterprise application, but a deliberate focus on decisions.  As Bill Fair (one of Fair Isaac's founders) once said, the key is to "grab the decision by the throat and not let go."  You must treat decisions, in other words, as separate entities addressed distinctly.
  
  
• automating and improving -- It's not enough to automate a decision; you must also improve it continuously.  Like automating bad processes, simply automating bad decisions results in little gain in anything except speed.  You should identify how you want to make a decision and then automate that improvement.  Improving a decision means not only working on every aspect of it to see where you can improve it, but also establishing processes and mechanisms for monitoring and constantly improving that decision over time.  This kind of ongoing improvement often offers the most value in enterprise decision management.
  
  
• operational business decisions -- Operational business decisions are those made in large volume every day.  They are clearly different from low-volume strategic decisions, such as where to open a new store or when to drop an unprofitable product line, that are rarely the same twice and simply don't happen that often.  Clearly, strategic decisions are important, but you aren't likely to automate them or try to make them in 'real time'.

  Note.  Some strategic decisions could be improved if the operational decisions derived from them are simulated and the results used to inform strategic decisions.  This method would count as using an EDM approach, albeit to consider a strategic decision's operational impact.

These workaday operational decisions are typically, although not always, part of interactions with associates.  These decisions have the highest volume and most time pressure of any in your business.  They are often hidden -- not specified explicitly, not managed, and not recognized as important.  You can probably think of many examples, including approve/decline, benefits eligibility, the next best offer to make a customer, authorization of a sale, fraud detection in a claim, issuance of a permit, and account application processing.  Typically, you must make these decisions in real time or near-real time.  Indeed, you might find that these decisions must be automated to achieve the requisite throughput and timeliness.

There's a gray area between strategic and operational decisions.  Decisions falling in this area are called 'tactical' decisions.  They determine how you manage processes and customers, such as decisions about which segments of a customer base should receive which offer.  You might support these decisions with EDM systems, but you're unlikely to automate them completely.

Operational decisions require the shortest decision latency, which is the time it takes to receive an alert, review the analysis, decide what action is required (if any) based on knowledge of the business, and take action. 

Precision, consistency, agility, speed, and cost -- Enterprise decision management has a wide-ranging set of effects that can be categorized as a combination of these five aspects of decision yield (which will be explained in more detail in a future instalment on "Introducing Decision Yield"):

• Precision -- how targeted a decision is
• Consistency -- across channels, geographies, and so on
• Agility -- time and cost to change
• Speed -- how fast a decision is made
• Cost -- how the decision affects cost

Key Features

Organizations have tried many different approaches to improve the way they make decisions.  Without trying to compare enterprise decision management to each approach in any detail, a few key features of the EDM approach are distinct:

• By focusing on decisions as distinct opportunities for improvement, enterprise decision management delivers solutions that are independent from applications and databases so that you can update decisions in one central location and use them in other systems, making those systems smart enough to run the business.
  
  
• By focusing on the power of having business users manage part of the decision directly rather than through a traditional IT process, enterprise decision management builds flexibility into decision processes and improves business agility.
  
  
• By using executable, predictive, and decision analytics to derive insight from data, an EDM approach means injecting intelligence into the transaction-processing stream for real-time decisions, not just after-the-fact analysis.
  
  
• With its focus on adaptive control and constant improvement, enterprise decision management requires a closed-loop capability for managing and improving decisions over time, which is made possible by the explicit focus on decisions and the use of technology and design approaches that support this focus.

These features are described in more detail in Smart (Enough) Systems.^[1]

[Sidebar-2.  Real-Time Decisions, Not Real-Time Modeling]

Characteristics of Decision-Making Problems

Realistically, not every decision lends itself to automation and management.  As noted previously, in most organizations, there's a spectrum from strategic decisions (where to build the next distribution center, which new product will make the most profit) to operational decisions (cross-sell offers, eligibility).  The right approach to supporting these decisions varies, too, from a focus on decision support for knowledge workers to a focus on decision automation, as shown in Figure 2.  Operational decisions are ideal for automation.  Tactical decisions, those falling in the gray area between strategic and operational decisions, might also be good automation candidates but usually for partial automation -- a guided script, for example -- rather than complete automation.

Figure 2.  The relative value of decision support and decision automation for different types of decisions

Strategic decisions might show results only over long time frames, suitable for reporting with standard business intelligence capabilities.  Performance management, meanwhile, is particularly good at tracking the effectiveness of tactical and operational decisions.  Operational and tactical decisions reward performance management because the time frames, and your time to respond, are much shorter, justifying the near-real-time supervision of results that's more typical of performance management.

[Sidebar-3.  Not Just Business Intelligence]

The next instalment will expand on what operational decisions are and will explain why and how so many of these decisions are hidden. 

References

[1] Neil Raden and James Taylor.  Smart (Enough) Systems:  How to Deliver Competitive Advantage by Automating Hidden Decisions.  Prentice-Hall (June 2007).  ISBN:  0132347962.  

# # #