This blog location is going to be retired and I will continue to blog at a new location - jtonedm.com. You can find all my posts there and can subscribe using the new feed feeds.feedburner.com/jtonedm.

Neil is blogging over on Intelligent Enterprise - check out www.intelligententerprise.com/blog/nraden.html

Having posted about Zementis - a company that allows you to deploy analytic models into the amazon coud - before I now see that Mathematica is getting in on this whole cloud thing. Personally I think that analytics and decisioning are ideal for operating in the cloud. Analytics take a lot of computing power when models are being developed, making the flexibility of cloud computing valuable. Decision management in the cloud means that any process, anywhere can connect to the cloud and get the questions answered it needs to operate effectively.

Interesting article on how the NHL used the power of personalization and targeting (or extreme personalization) to improve results. It does not talk about how they do this but clearly they have made things like the front cover choice for their catalog and key elements of their web presence decisions so they can make them differently for each person they target.

Some time ago I linked to an interesting looking research project and the results are now available - Development and Verification of Rule Based Systems - a Survey of Developers. Valentin did a nice job summarizing the results, comparing them to a previous study of more academic projects and drawing some distinctions between academic and commercial projects. Enjoy

Jim Sinur brought up an interesting point today when he blogged IBM, Microsoft, Oracle and SAP have bought Business Rule Technology. What’s up with that? The big players seem to be toying with business rules - there’s plenty of activity but not much understanding or commitment.

• SAP bought Yasu but until recently did not show much sign of “getting” the potential (though Sandy’s post on a recent SAP presentation sounds more hopeful)
• IBM is going to buy ILOG but we don’t have many details and the issue of how the WebSphere and Information Management/FileNet groups will integrate rules into their stories remains an open issue - there are folks at IBM who get it but concerns among watchers that the two groups will not make it work between them and only one piece (probably WebSphere) will end up with rules.
• Oracle bought Haley/Ruleburst but has made it clear that its platform products for BPM/SOA will continue to use their JESS-based rules implementation and Decision Service API. This raises an interesting question about Oracle’s future plans - another acquisition (Blaze Advisor, say) or development of their own (as Savvion did recently) as part of the Fusion plans .
• Microsoft meanwhile remains completely opaque with little bits of this and that as well as at least one company (InRule) with a perfect product should Microsoft decide to buy something.

So, will any of them “get it”? If so which one(s)? Guess we will have to wait and see but we live in interesting times…

I was pointed to a post today on the topic of customer service (Another Day, Another Customer-Service Nightmare on the EconoWhiner) that pointed out that companies

need to provide quality service and quality customer service if they’re going to survive an economic downturn as severe as this one?

Now I am not going to pick on AOL or even argue that his particular problem with AOL is the result of poor decision management (basic data management seems like the primary cause) - although making the billing decision a real decision not just a price look up would have allowed for discount rules based on problems.  What struck me was how many companies are listed in the comments! Clearly bad customer service is epidemic and, thanks to the internet, companies with bad customer service are going to be “outed”.

If you believe, as I do, that better customer experiences and putting your customers front-and-center will help you survive the recession, what can decision management do to help? First you could read my series of posts on using decision management to deliver a 21st Century Customer Experience. More specifically:

• You can make sure that your automated systems actually treat customers as people by making customer treatment decisions explicit and then putting your customer service people in charge of how those decisions get made.
• Take this approach to the self-service applications you have for your customers and you can make sure self-service helps rather than hinders
• You can also make sure your agents know how to treat your customers by providing them with decisions rather than data, targeted actions not general purpose lists etc. Your agents need to focus on the conversation and the customer, not on reading and analyzing customer data.
• Because customers respond to every decision you make as though it was personal and deliberate you can ensure that customer treatment decisions actually are by really personalizing your interactions
• You can also use decision management to make better churn and retention decisions and so retain customers - especially profitable ones!
• Using decision management to take control of customer treatment decisions helps make sure all your people and all your systems treat your customers the way you want to and so institutionalize great service

Check out my posts on Customer Experience for more thoughts

Predigy is a technology originally developed by Intelligent Results (founded in 2001) that was acquired by First Data in 2007. It was originally focused on the military (particularly on the analysis of unstructured data) but has subsequently moved into commercial applications. Predigy is now a decisioning platform with some applications in banking, collections, telecommunications and utilities. These applications are primarily in marketing, customer care, loyalty, retention, risk, fraud, collections and recovery. The intent of the platform is to support data-driven decisions throughout the customer lifecycle.

Predigy supports decision trees, predictive models and strategy design. It provides an offline business environment with some simulation and a production engine. This is all provided in a hosted environment and is web based as part of First Data’s environment. Predigy has four main components:

• Cluster
  An automated approach to finding natural groupings in portfolio accounts -Partitioning Around Medoid (an instance that best represents a cluster) is a proprietary clustering technique available only in Predigy. The use of the Medoid means that each cluster has an exemplar.
• Modeler
  Develop and deploy new models. Strong points are fast deployment, comparison tools and the ability to use directly in strategies. The models can include unstructured data and, like a growing range of modeling tools, provide a lot of automation of the leg work involved in developing models.
• Strategy
  Offline design of decision tree/model combinations - strategies - combined with the ability to use data to see business impact through simulation of strategies with various datasets.
• Production Engine

The product is increasingly integrated with First Data data stores and this means less data work for a typical FD customer. The base environment offers datasets, variables, formulas, samples, strategies, models, clusters and reports. Some interesting points:

• Variables can be explicit or calculated or even models - integration means that everything, even models, shows as a variable. Also supports formulaic clean up of data with extra variables -to fill in blanks for instance, transform and limit.
• Modeling is all automated, can use structured and unstructured data without issue. The automated unstructured analysis simply looks for strings and maps to outcomes - explicit entity identification in unstructured data is manual. The environment is aimed at modelers and focuses on automating base tasks like data cleaning and finding variables.
• Formulas allow you to define calculations about which you care in the tree. Formulae can be very complex and then built into decision trees. There is a function builder interface as well as java like language for modelers.
• Models can be created in other tools and use Predigy as the execution engine - the user simply pastes the model definition into a formula. PMML support is under development
• Sampling is support for building/testing, focusing in on a subset of population.
• Strategies are a decision tree environment. The tool tries to simplify the trees required by using models as characteristics and cuts. The user can specify costs and benefits of different actions on nodes for simulation/analysis and can drill into the data to see cost/benefit of the strategy. The decision making leaves on the trees have codes for either the customer’s own downstream system or the FD systems/call center/letter shop. The trees provide strategy code and action code to support Champion/Challenger.
• Clustering technology (PAM) is undirected model development - what are the groups in this dataset. Because medoids are real people the clusters can be described. The modeling environment is for targeting specific outcomes - directed analytics.
• The simulation tool flows sample data through the tree. The user can ask what would have happened if I had deployed this in the past. They can apply factors/constants/formulae to see what the outcomes would look like - target response rates for a given profit or return given a known response rate for instance.

Predigy is a nice, all-in-one decisioning platform squarely focused on the customer lifecycle, especially the customer credit lifecycle. It’s hosted platform and integration with First Data datasets are interesting capabilities. Currently it has no other rule formats- just decision trees, though this may change.

Last week I saw a post comparing Best Buy and Circuit City - one thriving and one going into bankruptcy - and it made me think about the role of decision management in Best Buy’s success. I have head Best Buy present various times an a number of elements of their successful customer-centricity strategy are decision-centric.

The first thing to note is that, overall, Best Buy focuses not on customer acquisition but on growing existing customers - especially since the best customers drive 4x the revenue of an average one. This focus on customer development and retention is common in decision-centric companies as existing customers generate lots of behavioral data that can be used to develop analytics and so improve decision-making. It also generates a great ROI.

One thing Best Buy has done is develop some very sophisticated personas. These are not based on judgmental criteria but the result of sophisticated analytics (I have discussed using analytics to improve personas before) and these models are used to asses the different demographic segments visiting each store so that they can be targeted with different store facilities and styles. Store layout and review for Best Buy is now very customer-centric rather than focused on store operations. This same insight into customer segmentation is used to change outbound marketing to a balance of mass-marketing, mass-personalized and truly personalized materials. This segmentation then is used not just as a decision support tool - to change the layout of stores - but also in an automated, decision management environment. This broad use of segmentation is another characteristics of decision-centric companies.

Best Buy uses analytics for more than just segmentation, however, additionally developing models to find products relevant to customers and to predict when they will be interesting. These purchase patterns over time are then used to drive direct marketing campaigns, increasingly across different channels. Some of these purchase patterns are long so Best Buy also uses the analytics to keep customers engaged in what can be a long sales cycle. For instance Home Theater is about 12 month cycle - 5 months dreaming, 5 months researching, 2 month active buying and Best Buy targets the 12 touches or so this involves. They also use these analytics to personalize the offers made through the loyalty program with targeted messaging, special offers in your areas of interest and so on. In the future they want to connect all this insight and apply more of their loyalty vision. This very fine-grained and targeted insight was described in a McKinsey concept of “cell-level insight” (see Capitalizing on customer insights - subscription required). The paper uses Best Buy as an example. One of the first things Best Buy did was hook up its revenue to individual customers, a key step to providing “cell-level” data. Once they had this “cell-level” insights Best Buy uses decision management to turn these insights into actions - you must embed these insights into “key decisions” to get value.  Best Buy, for instance, applies these insight to their direct marketing, offers and increasingly to the systems their staff use when interacting with a customer. Decision-centric companies are focused on using fine-grained analytics across channels and this focus on time-based analytics is where the more sophisticated are headed.

I believe that Best Buy is not successful only because it is customer-centric but also because it is decision-centric.

Recently, Ronan Bradley discussed the challenges for banks in the area of compliance, given the rapidly changing environment. He made three specific points with which I agree and that I think shows the value of a decision management approach for banks and others facing an unknown but difficult regulatory environment in the next year or two:

• “compliance will remain a major driver of IT projects in 2009″
  Compliance has always been a driver of decision management as the use of declarative, rules-based approaches and the explicit identification and management of decisions make it much easier for organizations to ensure and demonstrate compliance in those decisions. Opaque and poorly managed decisions are hard for regulators to evaluate and for companies to report on. Decision management eases the compliance burden significantly.
• “opportunistic responses to specific requirements”
  One of the values of decision management is increased agility in systems. Even if I don’t know what changes will be required, if I have adopted decision management then I will be better placed to find the decisions that must change and to make the right changes to those decisions. Decisions hidden in code buried away in applications or in process specifications are too hard to change fast.
• “much stronger focus on real time assessment of risk”
  Decision management allows for risk models - predictive analytic models of risk - to be applied in real-time in operational systems. Essentially every decision in every transaction uses the risk approach of the organization (as developed into its risk models) to help balance risk and reward for that transaction. Only decision management can realistically deliver this kind of in-line risk management.

So, to prepare for an uncertain regulatory future, adopt decision management now.

Savvion today announced it has released a Business Rules Management System. Now this may suprise you - after all Savvion is a Business Process Management vendor - but I think it is a sign of the growing recognition that decision management is important to business process management. Before this announcement Savvion was using Yasu’s product but, when that was acquired by SAP, they decided to develop their own.

While the new product is not the most sophisticated BRMS on the market, Savvion has done a nice job with the first version. They have provided an Eclipse-based environment that is integrated with their Eclipse-based BPMS tools but that creates independent projects. These can be deployed with a process or separately and so support both Savvion and non-Savvion applications. This is a key feature as decisions, and the rules within them, must often be reused across applications, channels and processes so restricting the decisions to execution within a process (as many BPMS vendors do) will not work. When integrated with business processes defined in Savvion then the rules and decisions consume the same information model - speeding development and easing integration. This new feature is in addition to their existing event rules where simple decisions can be defined with expression builder in the process modeler.

The core interface is what Savvion calls a decision table but that I consider a Rule Sheet:

• Each row is a rule
• Different columns represent the various variables used in conditions
• Each cell contains a specific condition against the variable of that column
• One or more actions are defined for each row

The rules can also be defined in Excel and then imported. The rules can be modified in Savvion’s portal allowing business users to maintain the rules in the same environment they monitor and manage their processes. The rules projects can be called as part of the process definition and as part of the presentation definition (for user interface elements built with Savvion). When working on processes or presentations you can navigate directly to the rules being used and manage them. Inferencing execution (using Rete) is supported as is a reasonable set of deployment options (stateless, stateful). All the process APIs can be executed by rules allowing for rules to drive the process nicely and everything is stored in the existing Savvion repository giving you versioning, access control etc. All in all a nice package, taking advantage of Savvion’s existing tools and repository.

I am pleased to see Savvion release this as I think it is important that BPM vendors get serious about decision management and that they realize (as Savvion did) that this is not the same as supporting process control rules. I really like the option to independently deploy the decisions as this helps people to manage process and rules as peers, another crucial step.

This is clearly part of a broader trend. SAP added Yasu to improve its rules capability in Netweaver, albeit only as part of the process definition. IBM is acquiring ILOG, presumeably allowing for independent decisions as well as improving the support for decisions inside WebSphere and FileNet. Pega has long had a rules-based approach to process management. With Savvion supporting not just rules in the process but independent decisions also I expect to see this trend continue.