Enterprise Decision Management - a Weblog

Market Consolidation

2008-07-29T03:49:08-07:00

IBM has announced their intention to buy ILOG. This is quite interesting for the BRMS market. Let me make a couple of comments on how this may change the dynamics.

After SAP acquired YASU last year, we all expected the beginning of a new acquisition cycle. Platform vendors have been shopping heavily for Business Process Management (BPM) and Business Intelligence (BI), skimming the marketplace to some extent...

A sample of BPM acquisitions:

Filenet has been acquired by IBM
Fuego has been acquired by Oracle (via BEA)
SAP has acquired a small BPM vendor, OutlookSoft
Staffware has been acquired by Tibco

A sample of BI acquisitions:

Oracle has acquired Hyperion
SAP has acquired Business Object
IBM has acquired Cognos

So, the YASU acquisition lead us to believe that Platform vendors were going to focus next on Business Rules Management Systems (BRMS). IBM just confirmed that trend. The size of the acquisition (one of two dominant players, Fair Isaac Blaze Advisor(r) being the other one of course) confirms what we have been advocating: "BRMS is a key component of the modern software architecture".

This is of course significant for us all, business rules evangelists and practitioners. IBM weight will now expose the technology -- and actually the entire discipline -- to a greater audience. Assuming that they are serious about getting into this business (which I do not doubt), this should boost adoption of BRMS at large.

What I find extremely exciting with this catalyst event is the prospect of elevating the debate to a whole new level. As BRMS will become a de-facto standard piece of enterprise software, Decision Management will spread as well. Why? This is actually a simple answer: as more people will understand the value of automating decisions, more people will also realize they need to make better decisions. Whether they drive those decisions from historical data using predictive analytics or they simulate and optimize those strategies in their labs, the Decision Management ecosystem will ignite the enterprise.

With the right determination on our part, and that of our followers, Decision Management may happen much sooner than anyone expected. That is a great prospect, for all of us.

Can we trust Business Users?

2008-07-07T23:58:44-07:00

-- Posted by Carole-Ann

The key premise of business rules management has been historically to empower business users. Since they own the business, they should be able to maintain the way they want to do business without being subject to IT interpretation and other backlog prioritization. This is the core value proposition. After 10 years of pitching this idea and making it real for hundreds of customers, one may wonder "can it really work?".

Given my responsibility of steering Blaze Advisor, you can guess that I have been a big fan of this idea from day 1. That said, regardless of the time and energy we have spent for over 10 years educating the marketplace, our work is not over yet... Let me illustrate.

I was recently in a friendly customer meeting. Several architects new to the product attended my presentation. I was surprised when one of them asked how business users could be trained to write performant rules. This is not a bad question when you know that some vendors put "lipstick on the pig". Providing a business user interface is far more creative than revamping the rule language with what looks like English or some business lingo. At least, it is if you want that exercise to be really successful.

The real secret for empowering business users is to separate the policies from the rule syntax, really. The approach I typically recommend is to design the rules maintenance application independently of what rules will be executed in the end. As a matter of fact, it does not even matter that a policy will eventually map to a rule, a portion of a rule, a function or a class or anything else. If you put those concepts forward, you will see yourself designing "business user" interface that appeals to rule developers, not business users. If you ignore those artifacts, you will likely focus on what those business users really care about: defining new mortgage or insurance products, drug interaction alerts or pricing schemes. Once you have collected the types of activity they need to perform, the complete screens they want to maintain (this is what we call the templates), you are then free to architect the business rules that support those and at that stage you (the IT person) can worry about performance and other important design considerations that should not be left on the business user's plate.

I often feel that this is underestimated. Too many rules specialists focus on rules that will get executed and not enough on the people that will maintain those rules, although this is where the success of the project resides.

Additional freebie for the Agile enterprise: when IT is working on the "rules content" of the templates, business users can start populating the rules repository. As long as the templates contain the display information, the business user interface is fully functional and therefore business rules can be authored. When those two parallel activities meet in time, you will be able to test the execution. With a fair amount of business rules available very early on, you can also adopt continuous testing for maximum agility.

So the answer is "yes" we can trust them but only if they can rely on IT to provide the right rules maintenance application. As always, it is teamwork.

5 tips for deploying predictive analytics with business rules

2008-05-31T22:23:13-07:00

Posted by Carole-Ann Matignon

A topic that passionates us at Fair Isaac is analytics deployment. We have spent 52 years figuring out how to build best of breed predictive models and how to make them available. I do fall in the same category... It keeps me awake at night. Not that I am worried about it, but I find exciting to keep pushing the challenge a step further.

You may have read many great posts from my colleagues on why it is a good idea, if not a necessity, to deploy predictive models and rules together. I would only quote one of those arguments here: time to deploy. We have heard many Financial Institutions complaining about how long it takes to deploy their current analytics. Almost half of the respondents to our survey said it takes 3-9 months or longer, over a year for many of them. Assuming those models address competitive pressure or a change in consumer habits / business conditions, it feels like a long time to endure once you have defined your new strategy.

With this context in mind, we have been working on many different ways to deploys those models in combination with business rules. The obvious reason we have explored this path is that business rules can be deployed in a very short time frame (technically in seconds, practically in hours given formal quality control processes). In one word: Agility.

We started many years ago with black box deployment that are made available to IT. For example, a Java deployment could then be integrated in a Java application. Many vendors are following this lead nowadays. We have since then adopted a complementary white box integration via PMML (predictive model markup language from Data Management Group). I would like to illustrate here a handful of lessons learned in that process.

Don't tie your predictive model life cycle to a given technology: Relying on standards such as PMML gives you independence on where you ultimately deploy your models. If you use a Java deployment out of your modeling environment, it is unlikely you will be able to deploy in COBOL or .NET natively. I recommend that you consider where you need to deploy today and anticipate as much as possible the additional environments you will need to support in the future.

Models should be used by business rules, not the other way around: Often you have a choice of where to put you business logic: eligibility rules may end up in the model itself, or they could reference the model (if your FICO score is too low, you will not qualify for a Jumbo loan). The key reason you should elect for the latter as much as possible is that business rules change and should be maintained by business users. Your scores may not change, or not at the same pace. Forcing your logic on the modeling environment forces the business to coordinate with modelers then IT before those changes get deployed into Production. You lose in agility what you gained in precision.

Consider your model's life cycle management as a process: There are still a fair amount of companies that exchange model definitions as a paper document with little or no traceability to the source. In today's world where governance is becoming paramount, it is important to start linking those artifacts with all necessary documentation. Besides future regulations, there is already value today in doing so. When models need to be refreshed, it helps to know where they came from and how they were developed (training data, exploratory process, etc.).

Don't assume that Modeling and Operational data models are the same: Modeling data is often prepared in a different way. The data model may have been flattened, data attributes may have been populated / filtered to avoid missing values that exist in real life, characteristics may have been pre-calculated, etc. Code developed to access the Modeling Data Model may therefore not be optimal or even executable reliably.

Do not underestimate the need for IT to "debug" the model code: Having a black box deployment prevents IT to access the model definition for debugging or runtime performance profiling session. Authorization mechanism and IP Protection tools can effectively protect the model definition if this is a concern. There are times when IT needs to get involved, typically during an emergency, so do not architect your solution to make it impossible.

With those tips in your hand, you can now question the technologies that you have selected and the current design for addressing decision management. How effectively can your system deploy predictive analytics and business rules together? Have you architected the process with enough flexibility?

Product Management, out of the box

2008-05-16T13:57:47-07:00

Posted by Carole-Ann Matignon

I have met many of you over the years at industry shows or in customer meetings, probably not all of you though, so an introduction may be in order. I am in charge of Product Management for the Decision Management tools at Fair Isaac, which means that I am responsible for elaborating the vision for our technology products, building the roadmap and making it happen with the collaboration of Product Development of course, and all other necessary functions. As you will soon discover, I have a real passion for decision management.

Several customers have insisted over time to have people like me reveal what we do, why we do it and how you will benefit from it. As this blog is changing hands, we have decided to finally give you this insider story directly from the Product team. It is an exciting opportunity to share my perspective and my excitement on our products, what new features are to be released and how we thought they would help you. I have also convinced by dear friend and colleague Carlos, head of Product Development, to join me on this blog.

Another area I would like to explore with you is our take on the direction of the marketplace, and our vision of how new technologies or new trends are impacting our industry,

Being in product management, I can't help myself... and have to ask for your feedback : -) That also includes anything else you would like me to blog about!

I am not afraid of speaking my mind but that does not mean that I am not also interested in a constructive argument: I encourage you to speak up and share your views on the blog, or for a private discussion send me an email at caroleannmatignon@fairisaac.com.

I look forward to our lively discussions.

Carole-Ann

Live from InterACT - It's a wrap

2008-04-30T12:15:33-07:00

(Posted by guest blogger, James Taylor)

Here's a quick summary of the posts from InterACT. Hope you enjoyed it.

Monday

Tuesday

Wednesday

This may be my last post on this blog - it's been over 250,000 page views, more than 800 posts and about 3 years, and it's been fun. Hope to "see" you all on my other blog - www.smartenoughsystems.com/wp

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

Live from InterACT - What do bankers worry about in the current crisis?

2008-04-30T11:04:00-07:00

(Posted by guest blogger, James Taylor)

Fair Isaac recently worked with TowerGroup, a well known analyst firm in financial services, to survey 108 executives at the top 50 U.S. banks and credit card issuers. The press release is here and there were a number of interesting results but two caught my eye:

The top investment area among respondents in the coming 12 months is in the area of analytics, with fully 75 percent planning such improvements
This is interesting as this is the opinion of people who already make huge investments in analytics. Clearly, even if you are already a very sophisticated analytic organization, analytics deserves (or perhaps even requires) an ongoing focus.
While roughly half of respondents currently have an enterprise-wide fraud solution in place in their organizations, those who do not are most daunted by the perceived complexity of implementing one
I was impressed that half already had an enterprise-wide solution but concerned about the other half. It's pretty clear that crooks will target any gap in your fraud defenses and so an enterprise-wide approach is really important.

Ted Iacobuzio of TowerGroup has a great quote in the press release:

No one doubts the seriousness of the current credit crisis, but it’s noteworthy that the largest financial institutions are more likely than others to characterize its impact as severe or worse. The fact that these larger institutions took on more risk in recent years and are feeling more pressure now that delinquencies are rising and defaults are increasing certainly has something to do with that perspective.

This reminded me of the presentation Joe Breeden gave yesterday in which he laid out the pattern of behavior that drove lenders, particularly large ones focused on how Wall Street sees their results, to put themselves on the path to the current crisis.

Ted is about to present and I will be blogging that over on the Smart (enough) Systems blog.

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

Live from InterACT - Tuesday Summary

2008-04-29T18:00:02-07:00

(Posted by guest blogger, James Taylor)

Here's a quick summary of the posts so far at InterACT. You can follow the full series over on the Smart (enough) Systems blog and subscribe by RSS or Email so you don't miss anything.

Monday

Tuesday

Back to the Smart (enough) Systems blog for the rest of the show, barring a last summary.

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

Live from InterACT - Bernhard Nann's Keynote

2008-04-29T08:41:02-07:00

(Posted by guest blogger, James Taylor)

Back to this blog for a session on Fair Isaac's product road map presented by Bernhard Nann, Fair Isaac's CTO. Fair Isaac is in the middle of a major evolution in their decision management solutions and in their platform for these applications. Today decision management solutions tend to be fragmented both by line of business and by customer life cycle - fraud, originations, credit cards, auto loans are all separate. Each of them is quite sophisticated but they are not connected, meaning that the decision process is discontinuous. Risk managers cannot manage the whole life cycle and criminals can use the gaps to target companies. This has to change he says.

The market for decision management applications has changed for a number of reasons.

Risk management increasingly crosses across silos and product lines, as does customer value assessment. Deposit, investment and brokerage accounts are being combined with credit accounts.
Chief Risk Officers are more common in financial services and more central to an overall risk strategy. Companies are talking about a corporate foundation for risk that can be pushed out to multiple countries and business lines while still taking advantage of local opportunities - control but flexible control with localized rules and models for local regulations, opportunities and customer preferences.
Banks and insurance companies are also looking to optimization and adaptive analytics, analytics that have a closed loop response, to improve results.
Finally, all of this comes with a greater pressure for transparency from Basel II and similar regulations - monitoring, reporting and notification must cover the whole life cycle and all products.

Technology too has been changed.

Distributed, component-based computing has become mainstream allowing a service-oriented architecture or SOA.
Disparate systems no longer require integration using batch transfers of data but take advantage of dynamic, more real-time integration.
The externalization of business logic has become mainstream with business rules management systems supporting business logic across the enterprise and across multiple decision management systems - many point solutions can share rules and decisions.

These two sets of trends and driving the new set of decision management systems Fair Isaac is delivering. These new applications will be connected so that decisions can be managed consistently across the life cycle and across products and so that preventative, not remedial, decisions can be made by applying what is learned at one point to another. These connected systems also improve visibility by consider customers across products, across accounts and over time. Integration of these systems into the corporate IT architecture is easier also, a critical issue as financial services organizations standardize on a common IT platform with more re-use, less data and logic replication. Lower total cost of ownership is critical both on the technical side and on the business side - the same code, the same data but also the same management of rules, same development of models. Organizations want to manage common logic and common models once and share them across multiple systems and use the same skills to customize their behavior in different countries. Fair Isaac feels it can do all this because they understand how to build decision management solutions, how to develop and use decision platform products like business rules management systems and analytic models. Their partnership with IBM allows them to use IBM's middleware - WebSphere - as a base. The new products will share a number of items:

Shared data model to ensure data can be shared effectively across decisions
A single business rules management system, Blaze Advisor, will manage the business logic across all the applications enabling decision logic, models and flows to be shared and reused
IBM's Enterprise Service Bus (ESB) makes integration of these decision management systems, and the components within them, easy and straightforward
A unified reporting environment provides transparency and understanding across the systems and customers
A single case management environment allows cases to move between groups and reflect what happens at each stage

Bernhard gave an overview but pointed a session that I will be blogging on the Smart (enough) Systems blog later for more details. Fair Isaac plans to roll out new versions of all its core applications over the coming years that support connected decisions and are built on a service-oriented component architecture.

Back to the Smart (enough) Systems blog for the next few sessions

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

Live from InterACT - Monday Summary

2008-04-28T17:43:55-07:00

Posed by guest blogger, James Taylor)

Here's a quick summary of the posts from today (and yesterday) at InterACT. You can follow the full series over on the Smart (enough) Systems blog and subscribe by RSS or Email so you don't miss anything.

More tomorrow - Bernard Nann's keynote on this blog, everything else over on the Smart (enough) Systems blog.

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

Live from InterACT - Mark Greene's Keynote

2008-04-28T09:00:00-07:00

(Posted by guest blogger, James Taylor)

As Mark is the CEO of Fair Isaac I thought I would blog this one over on this blog rather than on the Smart (enough) Systems blog where the rest are going to be.

Mark started off talking about the increased risk management issues prevalent in the markets today - the sub-prime crisis - and that have made it a tough year for financial services. Financial services companies are facing a range of old and new issues like rising risk, changing payment behaviors (what gets paid off first), organic growth, fraud, consolidation and its focus on operational efficiency, competitive pressure on profits and the old standby of compliance with new and old regulations. Managing risk in this environment is key and this, managing risk, is the theme of the conference.

A recent Tower Group study of banks found that the #1 technology imperative for coping with the current situation is improved analytics. This is clearly Fair Isaac's focus - new and improved analytic models and techniques. Fair Isaac scores, technology and approaches are critical in assessing risk, detecting fraud and managing clients. The challenges of the market are not lost on Fair Isaac, specifically whether the FICO score stood up to the sub-prime market. Fair Isaac feels that the score continued to do what it does but that, as always, it must be combined with other good lending practices like loan to value ratios. Part of the problem with the sub-prime crisis was an over-reliance on the score as a stand-alone item.

That said, clearly Fair Isaac could help in new ways beyond just educating companies on using scores. Innovation includes:

Forward-looking credit capacity score
Portfolio performance sensitivity
Portfolio valuation based on some of the same analytics as the Basel II work
Optimization, through the acquisition of Dash

Besides this research oriented stuff, Fair Isaac has also been working on FICO 08, a new FICO score. The plan was to improve the score particularly in its ability to predict risk in credit shoppers and sub-prime borrowers. The score should be available from the bureaus this summer.

Fair Isaac is also investing in its software platform for decision management. Using Blaze Advisor for business rules and Model Builder, Fair Isaac's analytic platform, the solution is based on the IBM Middleware platform. This was announced last year and new applications based on this will start coming towards the end of the year and will be discussed tomorrow in the keynote.

Research Exchange Partnership was another program announced last year to work on client's premises to solve tough problems. One of these was an optimization factory for HSBC where the research was to introduce uncertainty into the models - something I am sure Ian Ayres would appreciate given his focus on confidence intervals. The second was to help a large Canadian company to establish analytics to improve share of wallet by focusing on their most profitable customers and trading off profitability and attractiveness to customers. In particular new work on segmentation was critical - Action Segments is the name of the new product. Nominations for the next REP projects are open so it will be interesting to see what projects come up.

Mark closed with a nice little factoid - 10,000 people have attended InterACT over the years in the US, Europe, Japan and most recently China. Ian Ayres is up next and I will blog over on the Smart (enough) Systems blog for the next few sessions, including his.

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