Mike made an interesting comment in response to my recent post on the future of application development. He said:

Are business domain tools necessary to enable a non-trivial development role for businesspeople?

Like many good questions the answer is complicated - in a way, both yes and no. I do agree with Mike that a non-trivial role for business people requires that the environment in which they work - in which they “develop” or “maintain” system components or services - must be domain-specific. For instance, If a customer service manager is to “maintain the rules for prioritizing calls” then it mus be in the context of their business domain. It must talk about customers the way they do, describe the various queues or representatives the way they would, allow them to specify the rules in a way that makes sense to them (perhaps natural or near-natural language, perhaps a flowchart) and so on. Does this mean that such an environment can only be developed using a call-center domain-specific environment? No, I don’t believe it does.

The major business rules management systems have a variety of ways of making sure that the environment used by business users to maintain rules can be made to look and feel domain-specific. ILOG offers their Business Action Language, Haley a natural language interface, Fair Isaac a template-driven environment. Each allows the use of the business owner’s terminology and further allows the structure - the bones if you like - of the rules to be kept out of sight. As long as the objective is collaboration - IT builds some of the elements, the business builds other elements - then this works just fine. Once the environment is set up, the business is in the driving seat.

Where business domain specific configurations have a clear advantage is where an organization wants to take control of a standard process. Using something like Experian’s product (credit focused) or Chordiant’s decision management platform (CRM focused) makes it possible for a business user to jump right in with less technical collaboration initially.

I have written about this before and summarized it in the wiki under business user rule maintenance

• Forrester: The Forrester Blog For Application Development & Program Management Professionals Blog
  Interesting article-length post by Mike Gilpin on how process, service, event and information all hang together.
  (tags: mike gilpin Forrester process service information event)

Mike Gualtieri of Forrester had a blog post a few months back that I missed then but that he pointed out to me this week - What Is Your Future? In it he outlines two scenarios at either end of a continuum. One is that application development changes in incremental ways such that “The application development professional role remains largely unchanged, but they become more effective.” The other is a more extreme one that will “shift much of the responsibility of development to businesspeople”.

I believe that both scenarios will be widespread, but for different kinds of services or components. As the move to a component- or service-oriented architecture continues and as application development comes to mean composite application development, there will no longer be one application development scenario. Some components will remain the primary responsibility of people who look very like today’s application developers. Others will become the primary responsibility of business users or of business analysts. Others still might become the responsibility of analytic or statistical types.

For instance, services that manipulate technical components such as machines or sensors might remain in the hands of developers while those that make decisions as to what to do when a sensor tells you something specific would be developed by more business focused people. A business user might build a component that allows them to manipulate the information held about a customer that uses a more technical component that integrates the various data feeds from third party data sources with the internal databases involved as well as more analytical services developed to infer characteristics of customers from their past behavior. No one approach will work for all components.

I see this already, with the question of who is developing decision services become blurred with business, IT and analytic folks all working on them in different ways. Similarly, more business users are building their own portlets, reports and screens than ever before, even while most data and infrastructure services still require developers.

Clearly the development of different kinds of technology and the willingness of new generations of workers to use them and of management to allow them will be a big factor. The details of the resulting distribution of work are therefore hard to predict. What is clear is that technologies that empower the business and IT (and analytics) to collaborate will become more and more important and that organizations that get good at this kind of collaboration and at the development of decision services in this way will have an edge.

I have blogged about the good work Forrester is doing on dynamic business applications before and also about the role of business rules in building a digital business architecture and for concurrent business engineering. I will also be at the Forrester IT Forum so find me and say hi if you are there or subscribe to get the blog posts from the event as I write them.

Tony, over on the Decision Support Analytics blog, is running an interesting competition and the prizes have just improved - Neil and I offered a signed copy of our book to add to his list of potential prizes. You can check out the details here. I plan to write a couple of entries soon and I encourage any of you who blog to do likewise. Details of how to enter are here.

• Data Mining Research - www.dataminingblog.com: Data Mining Blogs: The Big List
  Nice long list of data mining blogs
  (tags: blog datamining data mining mine)
• Column 2 by Sandy Kemsley : Business Rules webinar replay
  Sandy’s nice business rules and business process webinar slides
  (tags: BPM BPMS BRMS BRM business rules process management sandy kemsley)

ILOG today announced a Scorecard modeler as an add-on for ILOG JRules® (which I first saw at DIALOG). As their press release says, this add-in allows customers to “incorporate statistical scorecard models directly into decision services” - a key tenet of enterprise decision mangement or EDM. ILOG is targeting financial institutions clearly but apparently also sees “a growing market for scorecard usage in other industries” as do we.

The Scorecard Modeler is an add-on to ILOG Rule Studio designed to allow users (typically business analysts) to directly create and manage predictive or additive scorecards. Scorecards are a well established way of implementing predictive analytic models for predicting risk and or customer behavior and integrating them into the rule studio allows rules and models to be used together.

The implementation in the JRules studio creates a new rules artifact that allows a business analyst or technical user to take a scoring model (built by a statistician using SAS or SPSS or something similar) and implement it on top of ILOG’s Business Action Language.

The scorecard allows scores to be assigned (scores can be formulas) based on attribute values and weighting to be specified. The scorecard allows a user to manage reason codes - they can specify a list of reason codes (name, code, description, priority) and then allocate them to scores. In addition it allows multiple reason codes and has some out of the box reasoning strategies to, for instance, handle ordering of codes, handle duplicates etc. Both the score returned and the reason codes can then be manipulated in rules subsequently. Overall most of the features you would want, including some nice nested scorecard support, is there.

There is currently no PMML import (something offered by Fair Isaac, Zementis and Pegasystems) but manual recreation of a scorecard is not a hugely time consuming step. In addition their point and click formula builder makes it easy to build calculated characteristics, a key issue in implementing scorecards. An early stage customer claims to have reduced total development time to about 1/3 of the previous time. The scorecard modeler is not currently available in their business user web interface.

While the scorecard is managed as a single artifact for versioning etc, logging and testing are done at the cell level (a cell in a scorecard typically represents a single rule).

ILOG made the observation that, even outside financial services, their customers are starting to understand the value of analytics in conjunction with rules and are excited about having scorecards where before they might have made do with rules and decision tables. As is often the case with business rules, the improved communication between groups that comes with a declarative metaphor is key.

Overall it seemed like a well thought out implementation, as you would expect with a development partner like Equifax who use analytics so extensively. As a firm believer in the use of rules and analytics together I am delighted to see another rules vendor add support for analytics, especially at a pretty sophisticated level. I am looking forward to seeing some PMML management in the future as well as some thought as to how to present analytic models to business users as part of their rules management environment.

Two of Gartner’s smartest analysts - Kurt Schlegel and Gareth Herschel (shameless plug) - just published an excellent little paper called “Business Intelligence and Decision Making“. This paper was one of Gartner’s Strategic Planning Assumptions and the (free) summary says:

A subset of organizations that seek a competitive advantage will evolve the primary role of their business intelligence and performance management initiatives to ensure that decision making is made a core competency across the company.

I particularly liked the paper’s focus on decisions at multiple levels - not just big, adhoc decisions but also the repetitive, operational decisions on which Neil and I focus - and the recognition that this is a stretch for those used to the current BI mindset. At only $95 the report is great value and you should buy it if you are not already a Gartner subscriber (and read it if you are).

Regular readers of this blog (and those who have read the book) will know how Neil and I think about this and how we like a focus on decisions and decision management rather than BI’s historical focus on data, aggregation and reporting.

The one thing I will give away from the report is it’s recommended reading list:

• Smart (Enough) Systems - our book - read about it here or buy it on amazon.com
• Competing on Analytics - Tom Davenport’s book (reviewed here)
• Super Crunchers - Ian Ayre’s very readable book on data mining (reviewed here)
• Moneyball - Michael Lewis’ classic on baseball and statistics

Enjoy the paper, take its recommendations (especially the one about buying our book).

I spoke to Seth Grimes last week about an article he was writing that just published on Intelligent Enterprise -What BI Practitioners Can Learn From Operations Research. As I was reading the article I also noticed a response over on Michael Trick’s OR blog -Business Intelligence and Operations Research. Both Seth’s article and Michael’s response are worth a read.

As Seth quotes me I think the “O” in “OR” is really important - OR is focused on operations and on the direct improvement of those operations - the improvement of decisions in operational processes and systems. BI has come to mean the support of more strategic or tactical decisions being taken by knowledge workers.
Thus, even if both arenas may be using, say, regression analysis, their focus is quite different.

Michael makes the point that “the field is not particularly excited about embracing a new name for our field.” but I think there is more to it than that. I am certainly not suggesting that OR needs a new name and I agree with Michael that “BI will inevitably use non-OR methods for some of their issues, so is rightly not “the same as” OR.” BI is not the same as OR and neither are the same as Enterprise Decision Management or EDM.

One of the reasons for using EDM as a phrase is to focus on the business issue - the management of decisions as though they are an enterprise asset - rather than on the tools/technologies/techniques being used (data mining, predictive analytics, business rules, optimization/OR, adaptive control). In the end the business needs to take operational decisions quicker, cheaper, more precisely and more consistently while being able to change them more easily. Any and all approaches are fair game in doing so.

Tom Jesionowski recently published an article on TDAN about The Prime Business Decision Loop. Tom had sent me a copy in advance to look over and I thought I would blog about it and publish my comments as I think he has made a valuable contribution to the discussion around decision making with the loop.

I like the four core processes he outlines - acquire, integrate, analyze, act - and I like the way it does not pre-suppose automated or manual decision making in its basic definitions. However, as Tom gets down into the details around “Integrate” I think he starts to assume that we are only worrying about manual decision making. For instance,  those building executable analytic models typically do not use data warehouse data and increasingly companies are also using adaptive analytics that respond to changing data automatically. Similarly when he starts talking about the act step I think he needs to think more about automation also, not just manual decision making. The act of making the right cross-sell, of approving an online application etc. is an automated action, not a manual one.

I really like the influence map and his discussion of personality in the communication around the loop, though I would add that the kind of automation involved is also relevant not just personality styles. Finally I would say that it is not enough to complete the loop “faster” than competitors, you must do it appreciably or noticeably faster - there must be some difference to the business. If your competitors do it in 3 days, doing it in 2.5 will not help much but doing it same day probably will, for instance.

So, while I like the loop overall, I worry that it assumes a manual decision making process. I wonder, for instance, what happens when the action is automated and the analyze step feeds not management processes but systems design.

Two articles caught my eye yesterday - Robert Nascenzi wrote an article “Real-Time Segmentation Levels the Playing Field” over on Destination CRM while Jeremy Nedelka wrote “The Ultimate Personalized Marketing” over on 1:1. Both articles focusing me in on what I have called “extreme personalization”. Jeremy’s article was a cute story about a school targeting applicants directly with personalized billboards and similar tactics. The campaign resonated with students because it “went out of its way to speak directly to them” and the article drew the (correct) conclusion that “in today’s world you have to create a deeper customer relationship”.

Now this is interesting but for many companies probably seems unreasonable. After all personalized billboards, for instance, won’t scale when you have huge numbers of prospective or actual customers. But analytic techniques can help out here. Moving on to Robert’s article we enter the realm of analytic segmentation. He gave the example of a company using advertising to bring in customers but then “Through no fault of their own, the wrong agents often make the wrong pitches with the wrong messaging to the wrong people”. He went on to say:

It’s time for companies to take all the knowledge they have about their customers, all their selling skill, and all the market research available on consumers as a whole — and apply it to brand-new prospects before the agent even picks up the phone.

and, obviously, continue to apply this throughout the rest of the conversation and indeed relationship.

Taken together these two mindsets, plus a focus on decisions, can deliver extreme personalization and here’s how:

1. Use analytics to understand and segment customers and prospects at a fine grained level
2. Establish a robust set of personalization information
3. Replace campaign thinking with decision thinking where the decision as to what letter to send this customer, which agent deals with this prospect, what offer to extend to this person, which local store to mention when talking to this customer is individual
4. Use the analytics and segmentation to make sure you make the right decision for each customer and use the personalization information to make it “personal”
5. Deliver these extremely personalized messages, offers, products to the right people across all your channels

This is a topic on which I have written before and one on which I am sure I will write again. The way to use analytics to deliver a great customer experience is not to present information to people, not to gather information into dashboards, but to deliver personalized, targeted decisions.