Enterprise Decision Management - a Weblog

(Posted by guest blogger, James Taylor)

The Performance Guys had a post today titled But is it Actionable? in which they discuss the general trend towards what is called pervasive Business Intelligence - that is, BI everywhere and for everyone. They note that the critical issue is not the user interface, browser-based or install or any of the other near-religious issues that come up when talking to BI vendors but the actionability of the insight delivered by the tool. As they say, the users need it to help them solve the problem in front of them.

This is a point of view with which I agree strongly. I would go on to make two additional observations, however. First, many of the people in a organization who need help to solve problems have neither the time nor the skills to use business intelligence tools well. They are not knowledge workers with complex, iterative problems but front-line staff with very specific, very repeatable problems. Second, any modern organization has lots of actions that are taken by systems rather than by people. Systems running in the back office, self-service applications for customers, automated links to suppliers and more. In neither case are BI or performance management tools going to be able to do much to help with action.

It is here that Enterprise Decision Management, EDM, comes into play. EDM is focused on the automation and improvement of decisions - making a system decide what action should be taken and then taking it - not on the support of manual decision making. As such it is both inherently actionable - as it is all about actions - and ideal for delivering those actions to front-line staff and systems. Instead of focusing on actionable insight, you could say that EDM focuses on insightful action.

As the Performance Guys noted, when thinking of BI or Performance Management tools it is not data, not information, but information that can be acted on that matters.Where EDM adds value is in those circumstances where the person seeing the insight may not understand or feel authorized to take the action or where these is no person to take action at all, such as in a web or self-service interaction. EDM is action-centric and uses the data you have to take more appropriate actions (by integrating predictive analytics into your business rules). If you think about the spectrum of BI from informative to actionable then EDM is the next step beyond actionable and into acted upon.

This is important to remember as far too many people, and vendors, are talking about operational BI as though improving performance, reducing latency and putting the same old BI tools in the hands of more users will get it done. I think that operational BI is (almost) an oxymoron as there is a yawning gap between how most modern organizations handle their operations (e.g. using systems) and the idea that there will always be someone to read the reports or use the analysis tools. The drivers for operational BI are real but the delivery requires new tools, new approaches, a new mindset. It requires EDM.

I often write on this topic on my various blogs and you might enjoy these posts and articles particularly:

• The future of BI had better include EDM
• The Three Dimensions of Intelligent Business Processes
• Customer 2.0, Operational BI and $20M
• TDWI says the most exciting part of Operational BI is Enterprise Decision Management (kinda)
• Decision Technologies and Active Data Warehousing
• Book Review: Competing on Analytics
• Those who know first, win - NOT!
• EDM and Business Activity Monitoring (BAM)
• Shifting Your CPM Into Action
• Gartner BI - wrap up
• Business Rules, Business Decisions, Intelligent Processes, Enterprise Decision Management
• The differences between BI and EDM
• Why ARE you still generating reports?

Finally you might enjoy this white paper I wrote called Smart-Enough Customer Decisions

Visit my Smart (Enough) Systems Blog(RSS) or my ebizQ blog (RSS). Buy the book or visit the companion wiki.

(Posted by guest blogger, James Taylor)

It seems to be customer interaction week (month) on my blogs. Today I finished reading an article that prompted this post - Service-Oriented Customer Interaction Management by Ashutosh Roy of eGain. There was a lot to like in the article but it seemed to me that a focus on the automation and management of decisions was a glaring omission.
The first line of the article is:

Brands are reinforced every day with thousands of well-executed interactions with valued clients

Now there's a sentiment with which I have to agree most strongly. Indeed, this reinforces the key premise of enterprise decision management or EDM - that little decisions add up and that a focus on the automation and improvement of these operational decisions is critical to the effective execution of a business. This is particularly true in the case of customer treatment decisions where an approach often called best next action is used to assess what will best develop the customer next time there is an interaction and then deliver the selected interaction across whatever channel makes sense.

Ashutosh goes on to make some good points about the value of a more component-based, service-oriented platform. In particular he talks about the problem that monolithic CRM systems tend to capture business process

in UI instead of workflow, knowledge, and reasoning

Again I think this hits the nail on the head. Capturing business process, assuming a broad definition of business process that is more than that which I can store in a BPMS, in its components is absolutely the right thing to do. The process and flow elements get stored as workflow and activities in a BPMS, the knowledge is stored as reusable and accessible content elements and reasoning... While Ashutosh does not really discuss reasoning explicitly, reasoning should be stored as business rules and analytic models that automate decisions. These decisions should be implemented as decision services (to fit with the other service-oriented components). Reasoning, in a customer interaction sense, must contain rules that are defined by regulation, by policy, by expertise and by the customer's preferences. It must also contain segmentation models so that different kinds of customers can be treated differently and it must use what is known about a customer to make useful predictions. Reasoning, in other words, is implemented using EDM. Regular readers will know that EDM also emphasizes adaptive control - the constant testing of champion approaches against potentially more effective challengers - to ensure that the way customers are treated is not allowed to become stagnant.

For this new architecture, he identifies 6 requirements:

• Open BPEL
• Open knowledge architecture
• Open, first-class data exchange interfaces
• Virtualized logical data views across repositories
• Componentized, lightweight UI widgets that can be deployed anywhere
• Packaged best-practice templates to bootstrap operation

To which I would add:

• Externalized, managed decision services
• A rule architecture that combines customer and company rules
• Adaptive Control to constantly test and refine new approaches
• Rapid deployment of executable analytics derived from all the data available

There's lost of posts on Customer Service on this blog and you might enjoy these particularly:

• Using EDM to improve the customer experience
• What's old is new - personalization is back
• Customer focus with EDM
• Using EDM to meet CRM challenges and use what you know about your customers
• EDM and Customer Centricity
• Live from InterACT: What's Next? The Best Next Action

I also posted a series of summary articles about using EDM to improve customer service over on my other blog. Check out these links:

• EDM and a 21st Century Customer Experience
• Using EDM to improve first call resolution
• Using EDM to keep loyalty where you want it
• Using EDM to personalize your business
• Using EDM to manage call center (and other) costs

Finally you might enjoy this white paper I wrote called Smart-Enough Customer Decisions

Visit my Smart (Enough) Systems Blog(RSS) or my ebizQ blog (RSS). Buy the book or visit the companion wiki.

(Posted by guest blogger, James Taylor)

Dale Wolf had a post this week titled 9 Steps to a Valued, Differentiated and Consistently Delivered Customer Experience. It was a very useful list and it seemed to me useful to map Enterprise Decision Management (EDM) into it to show how EDM can help you deliver the customer experience of your dreams.

1. Identify what customers value
   While this step is mostly about interacting with your customers, analyzing the behavior of customers on your website or in terms of the products and services they buy is an important consideration also. One of the best ways to tell what customers value is, after all, to determine what they are willing to pay for! Building analytic models of this will pay off both here and later when you automate.
2. Aggregate customers with similar values
   Classic use of analytics - segmentation. I would add that you should also aggregate customers with similar predictions for future behavior not just similar assessments of past values.
3. Determine value gaps for each cluster
   Classic marketing analysis, not much different for EDM-based marketing except that for each gap it would be useful to know when you might be able to decide to fill that gap.
4. Prioritize gaps you can best fill
   Again the only addition would be to include an assessment of your ability to act at the right moment in your prioritization.
5. Determine tools to deliver value
   Which channels, which interaction points, which decision points.
6. Eliminate all non-value adding messages, tools and channels
   And make sure that all those you keep are consistent and deliver the same decisions consistently.
7. Improve and automate critical processes
   And the decisions within those processes. Deciding how and when to act to improve customer experience is critical. EDM is focused on those moments of decision within your customer service and marketing processes. Don't forget to consider all the decisions in all the processes that might impact a customer experience so you can reuse the right decision across them all.
8. Select contextually relevant strategies to communicate value
   An important note - just because you are trying to take a consistent decision does not mean that the context in which you are taking the decision is not both an input to the decision and relevant to how you act on the decision. You want to make the same cross-sell offer, perhaps, but would do so differently in different channels.
9. Measure continuously against goals
   And implement adaptive control to try alternative approaches as you go. You may think you have the best approach but not only should you monitor this approach, you should have some way to try alternative challengers all the time.

Interestingly enough I posted a series of articles about using EDM to improve customer service over on my other blog. Check out these links:

• EDM and a 21st Century Customer Experience
• Using EDM to improve first call resolution
• Using EDM to keep loyalty where you want it
• Using EDM to personalize your business
• Using EDM to manage call center (and other) costs

Visit my Smart (Enough) Systems Blog(RSS) or my ebizQ blog (RSS). Buy the book or visit the companion wiki.

(Posted by guest blogger, James Taylor)

For those of you who don't get DM Review magazine in your mailbox, you might want to check out a great new DM Review columnist - me!

The column is called Smart (Enough) Business and the first one is available here - What is a Smart (Enough) Business Anyway? (free registration required)

Visit my Smart (Enough) Systems Blog(RSS) or my ebizQ blog (RSS). Buy the book or visit the companion wiki.

(Posted by guest blogger, James Taylor)

Adam Sarner at Gartner published Five Steps to Successful Event-Triggered Marketing last week. The abstract of this paper says:

Successful event-triggered marketing is a process of identification, categorization, monitoring, optimizing and executing. Marketers that do this right will see their marketing messages receive up to five times the response rate of nontargeted push messages.

It seemed to me that this was a good prompt for a quick reminder on how EDM - Enterprise Decision Management - can be used to deliver on the promises of event-based marketing. Taking Adam's five stages:

• Identification of the events that matter requires an understanding both of the rules of your business and your historical data. Considering events that can be described in terms of rules (customer's total balance exceeds $X) and analytics (customer's retention risk exceeds Z%) gives you a rich yet actionable set of events. You won't be able to monitor all of them but this will give you a nice superset.
• Categorizing the events to help decide which ones are important is critical to delivering something practical so use predictive analytics when categorizing. Predictive analytics, with their view into the future, will help to identify leading indicators where other forms of analytics will tend to be backward looking. Analytics for segmentation can also help determine when particular events matter to particular sub-groups of your population.
• Don't forget when monitoring that you must consider all channels. Using a common decision service to decide which events have happened enables you to detect these events across all channels thanks to decision externalization.
• Use adaptive control, champion/challenger testing and optimization to make sure that you are making the best offer for each event for each customer. What is best will change over time as your customers, competitors and the overall market change. Stay focused on the micro decisions of how to respond to a specific event for a specific customer to make the most of event-based marketing.
• Consider using more decision services to handle execution so that you can work across channels. A managed decision service can determine the next best action to take and then decide which channel to use to communicate with that particular customer using a combination of effectiveness and preference.

EDM is a very effective tool in your toolkit when implementing event-based marketing. Here are some other posts on marketing you might enjoy also:

• What's old is new - personalization is back
• Is your system smart enough to make your customers feel good?
• Customer focus with EDM
• Using EDM to improve marketing operations
• Data Mining, Predictive Analytics and Markets of 1
• A home page is a decision
• Hits and Niches
• Here's a way to have 1:1 communication and scale too
• Event-driven and decision-centric

Visit my Smart (Enough) Systems Blog(RSS) or my ebizQ blog (RSS). Buy the book or visit the companion wiki.

(Posted by guest blogger, James Taylor)

I was pointed to a new site this week - Gonzo Banker - and in particular to this article on Workflow: The Killer App. While I enjoyed the article, I do think that the example used shows the weakness of a workflow-only solution. The article begins with some nice definitions of the two BPMs:

The implementation of workflow technologies is the foundation of business process management. Measuring the effect of implemented workflow is a key aspect of business performance management.

I would add that adding intelligence to workflow is the purpose of enterprise decision management and just as important as the two BPMs. This is clear even from the example used to illustrate the power of bpm.

Example: Tom, a CSR, requests to rebate an NSF charge that requires approval from his branch manager, Susan. The required data is the account number, the date of the fee, the amount to be rebated, and the reason for the rebate (required for performance management, not process execution).

In the old days, Tom might draft an email, send it to Susan and ask her to approve and forward to the Deposit Services email group. In the really old days, Tom printed a form and had Susan sign it and route/fax it. Either way, the information on the form could be incomplete. By using a workflow engine, Tom cannot submit the request until all data is entered.

I would look at this differently because there is a decision in this sequence - should a rebate of the NSF charge be given. The assumption in a workflow-based solution is that this must be made by a person other than the person gathering the data. But the person to whom it is being referred is not going to collect any more data which means that all the information required to make the decision is gathered at the first point of contact. Instead of thinking about workflow, think about the decision:

• Use business rules to make sure that the necessary data has been collected.
• Automated a decision to rebate or not based on the bank's rules, the relevant regulations and (possibly) predictive models for fraud risk.
• This decision service can use data without exposing it to the CSR and so privacy rules are easily enforced.
• Immediately give the CSR the answer and reasons for it.
• Collect the rules and model execution data so that the supervisor can review them later if she wants.

In this case there is no need for workflow - why move work around when you can decide and then act? Taking an EDM approach in this case means that the supervisor can spend her time on more productive activities than rubber-stamping approvals and that the CSR can give a customer immediate feedback - no open ticket, no customer waiting, no workflow!

Banking, being highly regulated and very risk-centric is full of these kinds of decisions - ones that used to be referred to a manager but that could, in fact, be automated with the right approach - EDM. Here are some other posts on banking and financial services that will give you some more examples of what I mean:

• 8 ways to improve captive finance with EDM
• Better Bank Branches with EDM
• EDM-Enabled Web Self Service Better be Top of Mind for Financial Services Firms
• Using EDM to respond to trends in Financial Services
• Why Banks need to focus on Growth Decisions
• Using decisioning to build the bank of the future
• A banking story
• Here's a quick way to get more automatic loan decisions
• Self-Service in Banking

Visit my Smart (Enough) Systems Blog(RSS) or my ebizQ blog (RSS). Buy the book or visit the companion wiki.

If it seems as though your world is spinning faster, you’re not alone. Credit managers and insurers throughout Europe and around the globe face rising levels of uncertainty, increasing regulations and higher levels of demand from consumers. The risks are greater, the profits more elusive and the competitive stakes more daunting than ever.

Join us at InterACT, the Smarter Decisions Conference, 3�6 March in Vienna, for a refreshing new view of success. Now in its 32nd year, InterACT brings together a wealth of perspectives�credit managers, insurers, regulators, industry experts and Fair Isaac advisors�to help you exercise greater control over your business.

InterACT provides unparalleled insight into the latest technologies and strategies that can help you make smarter decisions across the customer lifecycle. That’s why Fair Isaac’s InterACT conference has been growing in attendance each year, and has become a can’t-miss event for hundreds of businesses.

Download this brochure to learn more about our upcoming event in Vienna.

Download InterACT Vienna Brochure (pdf)

Or click here for more information and to register: www.FairIsaac.com/InterACT

As is our regular practice, we shall very soon publish special discount codes that will give you privileged access to lower prices at the event. And be on the lookout for details of our event in San Francisco, too.

(Posted by guest Blogger, Gib Bassett)

Gartner recently issued two “Predicts: 2008” reports highlighting their expectations for Enterprise Architecture and the Business Process Management Suites (BPMS) Market. Both recognize noteworthy and highly complementary trends reflecting the value of Enterprise Decision Management (EDM), particularly Business Rules Management Systems (BRMSs) supporting EDM principles.

In the most recent report, issued December 11, 2007, and titled “Business Process Management Alters Business and IT Collaboration,” Gartner predicts that “BPMS Subsumes the Business Rule Market,” primarily due to the fact a baseline rules engine is a required component within a BPMS:

“Strategic Planning Assumption: By 2012, the BPMS technology market will subsume greater than 80% of the pure-play business rule technology sector (for example, stand-alone business rule engines).”

Readers should not mistake this statement as implying the market for Business Rules Management Systems will be subsumed, which is not likely simply because a BRMS has applications and value beyond the scope of process automation, and supports an EDM approach to decision automation, a caveat noted in the report:

“The BPMS technology market will not subsume 100 percent of the business rule sector due to the fact that rule engines are used in many other technologies outside of BPMS, such as portal management technology and traditional applications.”

This section of the report also describes conversations Gartner has had with its clients, who claim to find it difficult to model processes and rules together. This might make sense in a process-centric context, but when considering business rules on an enterprise basis that may be tied to the execution of analytic models and other applications, it is wise to consider a dedicated BRMS. Again, Gartner recognizes this important distinction:

“For BPI (business process integration) projects that need extremely sophisticated rule-processing capabilities now, buy a BRMS. Don’t wait for BPMS providers to catch up to BRMS functionality if they are currently lacking it.”

Gartner goes on to make two final points worth mentioning:

“Due to the increasing need to model rules and business processes simultaneously, it is going to be increasingly difficult to maintain separate rules and process repositories.”

And

“Enterprises need to understand the consequences of not managing rules and processes with a shared metamodel, as well as the effects this will have on their ability to create dynamic business models.”

The latter of these quotations is especially noteworthy, as it resembles Forrester’s prediction for Dynamic Business Applications which feature rich business rules management capabilities to enable the adaptive and agile composite applications of the future. Gartner’s commentary seems driven by what it sees as limitations in many offerings today used in process and rules integrating projects. BRMSs compliant with EDM concepts can overcome much of these problems via deployment of rules and analytics as Decision Services.

This leads to the other report, issued November 12, 2007, and titled “Emerging Trends Force a Clearer and Deeper Focus on Enterprise Architecture.” In this “Predicts” 2008” paper, Gartner proclaims:

“By 2010, the primary focus of technology architecture will shift from defining product standards to identifying and describing shared and repeatable technical services.”

You can be assured these services will include Decision Services that will have the additional benefits of addressing these points discussed later in the report:

“…mounting legal and compliance pressures will require the EA teams to define methods to support organizational needs for the transparency of information assets, because data must be protected, validated and located as it moves from upstream transaction processing systems to downstream analytic environments.”

BRMSs compliant with EDM concepts include both enterprise-grade governance and analytic deployment capabilities. All of which foretells a promising future for EDM.