It’s the end of day 1 of the Business Rules Forum/Enterprise Decision Management Summit and time to write a wrap up post for the day – no live blogging today as I have too much on as track chair to sit behind my keyboard!
Today I got to attend Jim Sinur’s keynote and sessions from Roger Ahern of Experian, David Wilson of John Deere, Stephanie Alsbrooks and Kartheek Veeravalli of ThinkCash Financial and Sundar Vallianayagam of Jarus Technologies. Here goes with the summary:
The Decision Dilemma: Making Better Decisions in the Face of Uncertainty
Jim Sinur started the day off with a presentation on Gartner’s perspective around pattern-based strategies. Arguing that things will never be the same after the recent meltdown and that change is the new constant for companies, Jim made a pitch for pattern-based strategies. Companies must find a way both to be proactive to change that might reasonably be expected as well as rapidly reactive in identifying and responding to unexpected change. This requires a focus on leading indicators, especially those outside the organization, as well as ways to make people, process and technology change happen quickly. A pattern-based strategy that seeks, models and adapts suitable patterns is essential. Business rules, business process, event processing, streaming data are all key to this approach.
Organizations must become pattern seeking, performance-driven and transparent. They must also seek what Jim called the Optempo advantage – an ability to improve their competitive rhythm and increase their rate of operational adaption. He identified 4 key disciplines:
- Seek patterns that will matter to your organization, especially strategic opportunities for innovation
- Become a performance-driven culture, one that measures and improves
- Match the pace of change, your ability to change, to a business purpose
- Become transparent about why and how you do things
This requires organizations that use predictive analytics to look into the future, social media to understand what people are saying, complex event processing/decision platforms to handle real-time response, business planning tools to be specific about the kind of organization you are running and BPM/business rules/BAM to ensure systems are adaptive to change.
As usual Jim put a lot of material out there and asked a lot of interesting questions. With the keynote done it was time to go into the decisioning sessions.
Aligning Decision-Making Improvement Initiatives with Business Goals and Objectives
Experian has been working on helping companies automate the decisions in the credit lifecycle for a while and has identified 19 different decisions that can be usefully automated. The value of this automation comes in terms of improved accuracy, agility, speed, efficiency and consistency. Roger pointed out that a bank with just 500,000 customers probably makes more than 20,000 credit lifecycle decisions every day. With the difference between a good credit decision, that might make a $150 in profit, and a bad one, that might lose $1,000 or much more in fraud, this becomes very significant quickly – little decisions add up, as I have said before. When a company like HSBC makes 50Bn decisions a year, it becomes even clearer…
Experian works with a lot of banks and, like me, really tries to focus customers on mapping decisions to the metrics that those decisions influence (after all metrics are what most executives are accountable for). Given that rules and analytics, decisioning technology, can improve most decisions by 10-15% in their experience it is not normally hard to see the likely benefit in terms of business measures where decisioning can move the dial. Nevertheless, companies must prioritize decisions y their impact.
Experian sees decisioning having several pieces – a data and attribute engine that manages data and calculated attributes; a decision engine that handles scoring, policy and decisioning; and an optimization tool to create optimal approaches. Decisioning can and should be used to enhance existing applications and processes, enhancing them with advanced decisioning.
A solid presentaiton on the value of decision management and on the adoption of decisioning,
From Raw Data to Proactive Decisions: Using Business Rules for Condition Based Maintenance
From financial services credit decisions to proactive maintenance of earth moving equipment was quite a shift but it showed the value of the rules approach in a broad range of problem domains. David talked about a new product being rolled out by John Deere to take all their maintenance, testing and telemetrics data and turn it into actual recommendations – suggested work orders – for their distributors. The idea is to take the know-how inside John Deere and put it to work in the field by making sure that all machines sold by John Deere are given the preventative maintenance their need.
The decision engine John Deere has built pulls data from multiple sources and uses it not to make pretty pictures but to generate concrete recommendations for dealers – things they can and should do right now. The previous approach of just making the data available failed what I call the “so what” test – no-one knew what to do with the information. With the new system the data that is available is used to drive specific, actionable recommendations for dealers. Instead of having to wade through lots of data, and lots of paper to understand the data, the dealer gets concrete next steps for each piece of equipment. Building the engine that recommends the right next step should mean that more preventative maintenance is done, making machines do more for less for customers and reducing warranty work for John Deere – a win win.
The original approach was to try and capture the knowledge in Word and then Excel but the project only really got underway when they adopted a business rules management system to manage the logic. With this system the business users are able to design the rules that are needed while the IT folks manage the detailed implementation. The initial development is in pilot now and the ability to rapidly revise and redeploy rules has been as critical as the original improvement in development. The changes to the rules that “real life” requires are significant, even when experts are writing the rules, and ongoing maintenance and updates will be critical.
John Deere expects to gain real competitive advantage out of this system – it will help ensure that their dealers do the best possible job helping companies keep their equipment running effectively reducing down time for the customer and warranty costs for John Deere. The release is soon and personally i think it is going to be wildly successful – it smells like one of those great rules ideas.
Addressing the Challenges of Developing a Universal Decision Management Platform in an Emerging Company
This presentation was about a previous project, nothing to do with the presenters’ current employer, but was a great discussion of how to develop a coherent decisioning strategy and platform. The project came about when a new auto lender was formed that wanted to target all possible auto borrowers (prime to sub prime) while managing risk in an innovative way and delivering a better dealer experience. With a plan to develop a cost-effective, customer-centric consumer platform with flexible processes and a rapid time to market, the team made some interesting choices.
First and foremost they decided to buy the common processes – those where little differentiation was possible – and build a unique and differentiating decisioning approach. This would give them a competitive edge without requiring a completely unique platform. They also outsourced the development of IT infrastructure and workflow while retaining critical decisioning know-how in house. This let them develop a platform for processing loans quickly while still providing differentiation. The decisioning originally focused on origination before expanding into servicing and funding. The initial releases involved manual review but rapidly evolved to fully automated decisioning with random manual audits.
Segmentation of prospects/customers to manage risk effectively and champion/challenger adaptive control were both critical to rapidly and effectively evolving a unique risk management approach. Simulation was widely used to understand the impact of a potential change and this was combined with the champion/challenger approach to ensure the business understood the impact of any change being considered. The combination of predictive analytics for risk assessment, optimization for pricing and business rules for policies and decisioning resulted in a very sophisticated decisioning platform. Most decisions were real-time and 100% automated and the business and IT groups were able to collaborate on the definition of these decisions.
Automating Commercial Underwriting Using Business Rules
Last session of the day was on commercial underwriting using business rules and predictive analytics. The platform that was built in this way handled all the key decisions – eligibility, underwriting, quoting and binding and issuing the policy. A business rules management system allowed business analysts to manage the business rules while also providing a mechanism for quick and accurate deployment of predictive analytic models developed to predict risk. Critical to the project were effective rule harvesting (what rules are involved) and rule architecture. The team had IT work on the actions that rules could take and this allowed the business analysts to specify the conditions for each rule while picking one of the allowed action. This simplification allowed less technical users to participate in writing and maintaining rules.
The combination of predictive analytics with this rules-based environment used PMML and took advantage of the flexibility of a rules-based engine to allow rapid model deployment. This combination allowed for what the customer called “surgical pricing” and the retention of the risks the company wanted at a price that made sense. The project won an Innovate award from Insurance Networking News.
A classic rules and analytics story in insurance.
Here are some posts from the conference – Sandy Kemsley, Paul Vincent and Eric Charpentier all posted great information:
- The Decision Dilemma #brf
- BPM, Collaboration and Social Networking #brf
- Process Notations #brf
- Collecting, Connecting and Correcting the BPM Dots #brf
- BRF09: Jim Sinur on Business / Event Patterns
- #BRF Making Better Decisions in the Face of Uncertainty Keynote
- #BRF BPM, Collaboration and Social Networking Presentation
- #BRF Business Rules enhance agility in BPM presentation
- #BRF Enabling Effective Business – IT Collaboration Presentation
- #BRF Enabling Effective Business – IT Collaboration Presentation
- #BRF An Evolutionary Perspective of BRMS Presentation
- #BRF Vendor Panel: BRMS at a crossroad?
If you are following along on twitter – check out #brf as a tag.
Comments on this entry are closed.