Analytics & Optimization Ch-ch-ch-changing….

Mar292006

Brad Appleton posted a great little piece on Unchangeable Rules of Software Change - Redux. He makes some great points in the article and I am going to shamelessly steal some. He identifies three pitfalls that are very pithy: Pitfall #1: No scope change-management so change overwhelms the team Pitfall #2: React to pitfall #1 by trying to prevent scope changes - "if only the requirements would stop changing before a single line of code is written, then the schedule wouldn't keep fluctuating so much" Pitfall #3: Requirements Analysis Paralysis to try and prevent problems His summary - "It's those darn users/customers/analysts that don't know what exactly they want at the outset. It's all their fault!" -  is perfect. This is the basic driver for many business rules implementations - those focused on rules that change often (one of the four types of justifications for business rules). Brad then outlines some rules that I also really liked: Rule #0: Change is Inevitable! Rule #1: Resistance is Futile! - There isn't a darn thing you...

2 Comments

Analytics & Optimization Boosting Call Center Performance with EDM

Mar282006

Another great McKinsey on IT article - Using IT to boost call-center performance - caught my eye. One quote in particular leaped out:The next wave of improvements will probably come from using technology to automate contact with callers, to help front line staff resolve calls, and to handle call volumes more efficiently.If we think about how one could apply EDM in a call center, the first two issues come to the fore - automating contact with callers and helping front line staff resolve calls. These kinds of high-volume, automated transactions are the ideal target for EDM projects. Part of completing these transactions is to make a decision immediately, even when that decision must follow complex rules (state and local regulations that apply to this particular customer, retention policies, refund rules etc) and apply risk-based or opportunity-based criteria (estimated lifetime value of a customer, credit worthiness of a customer etc). Applying EDM in these circumstances means: Identifying the decisions that enable customers to self-server and call center staff...

Leave a comment

Analytics & Optimization Railroads, logistics and EDM

Mar272006

Interesting article in InformationWeek today - U.S. Railroads Rolling Out High-Tech Logistics - that made me think about applying EDM in railroads or other logistic problems. As I read the article I started thinking about the operational systems that could be impacted by an EDM approach. Developing operational systems that manage the real-time routing of cargo using business rules (to implement expert know how and contractual implications) and predictive analytics (to predict bottlenecks, risks of delay etc). Some other logistics companies are doing this and some Telcos are using rules to reconfigure the network when they have failure, something that is comparable. Using business rules to be proactive when service levels are not going to be reached or shipments are going to be delayed and calculating the - I know some Telcos doing this. Deploying business rules to laptops to ensure that safety and security inspections are rigorous and meet the federal, state and company standards established - some ship inspections use this approach already. Using business...

Leave a comment

Analytics & Optimization Out of the labs – business rules already empower nonprogrammers

Mar232006

This article in Computerworld struck me as interesting - In the Labs: Automatic Code Generators. The subhead was "Even nonprogrammers can 'program' with these tools". There are three issues with the premise of this article. Firstly why would you wait for things in the "labs" when there are technologies out there already empowering nonprogrammers? Secondly if I generate code, what do I debug when it does not work? Thirdly, who says nonprogrammers want to program? So, first things first. Business Rules Management Systems (BRMS) or Business Rules Engines (BREs) have been around for years and already work. They are proven in huge installations (10s or 100s of thousands of rules) and have installations (like the one at Egg) where business users have made every change to the application logic for years without IT and without bringing down the system for maintenance. Why wait for some new technology to come out of the labs when you could buy a proven product today? Second issue. If I generate code what do I debug? If a user builds a model or draws...

Leave a comment

Analytics & Optimization Business and IT collaboration

Mar212006

Randy Cronk asked an interesting question the other day - he asked me to cite some examples of making it easier for business and IT to collaborate (see his comments on this post). This made me think of a recent ADT article - I blogged about it here - in which a couple of related points were made: Giving that sort of power to users, even savvy ones, presents challenges. "If you give non-technical users the ability to use declarations to change business logic within an application you have to be [careful] that you're not going to cause performance issues." Eventually, users may be able to write and maintain business rules in ways beyond the examples here, but for for the time being, that's still what Forrester's Rymer calls "the new frontier." It also reminded me of some comments I heard amongst some senior IT staff from healthcare payers - they were not convinced that business rules would allow business users to "do it themselves" or even that this was a good idea. Now the common thread here is the need for collaboration and the...

3 Comments

Analytics & Optimization AIG Trade Credit – a case study

Mar212006

Digging around in some old case studies and I found this one from AIG Trade Credit. This one was an interesting one for a couple of reasons. Firstly its a great example of how business rules can help IT and the business collaborate when designed right and secondly because of its focus both on agility (responding quickly to market changing events) and on costs savings (allowing more IT dollars to be spent on development and fewer on maintenance). A couple of quotes from the case study stand out in this respect:Above all, Blaze Advisor allows business users and analysts who understand credit risk to write and audit changes to the system’s credit models without resorting to a programming language or relying on IT resources for changes made during application release cycles. [The CIO] says “We see business rules as an essential technology that will continue to be fundamental to our agility in reacting to external market forces and rapidly executing our business strategies."AIG Trade Credit achieved this by having the IT folks lead the development of a system...

Leave a comment

Analytics & Optimization Trends in Financial Services – the decision-centric institution cometh!

Mar202006

Duffie Brunson wrote an interesting piece on b-eye-network recently - Dominant Trends in Financial Services - based on Knightsbridge Solutions annual white paper. I am not going to try and summarize it all but a couple of points did leap out at me Duffie talked about a compliance calm with some of the big drivers like BASEL II getting held up for US banks at least. This has allowed banks to take stock of their compliance investments and work on building a better data infrastructure for the future. It should be noted, however, that international financial institutions are getting less of a breather and that compliance still requires compliant processes and decisions, not just data. There's a lot of value in EDM when it comes to compliance, with more articles etc here. One trend is Enterprise level BI, especially for risk data at an enterprise level. Clearly the growth in availability of enterprise-wide data increases the odds, and value, of enterprise-wide decisions. Having risk data across the enterprise enables risk-based decisions to be enterprise-wide and...

Leave a comment

Analytics & Optimization Enterprise Risk Management and Enterprise Decision Management

Mar162006

Interesting article by Steve Minksy in ebizq today - The Dos and Don’ts of Enterprise Risk Management.  Enterprise Risk Management is a key objective in many organizations that adopt EDM and not just because there is only a one letter difference in the abbreviations. Steve founded RulesPower, a company acquired by Fair Isaac last year with an eponymous product in the process of being integrated into Blaze Advisor. Steve offers some great advice on "how to" and outlines some nice criteria for ERM Technology. Clearly the combination of business rules (for compliance, process control) and analytics (for risk prediction) is key here. I like his last comment the best though:The greatest rewards come from Enterprise Risk Management when used to affect the way all our decisions are made. The key is in knowing what to do, and what not do to. Bolstered by the right technology, smart companies really can turn risk into gold.This is the essence of EDM applied to risk - build/buy systems that focus on decisions and how they are made. Get ALL the operational...

1 Comment