Predictive Analytics World, February 16-17, 2010 at the Palace Hotel in San Francisco is turning into the biggest yet. I am going to be speaking on analytic journeys and giving a workshop on putting predictive analytics to work and there are some great keynotes from Andreas Weigen (ex-amazon.com), Kim Larsen (Charles Schwab) and conference chair Eric Siegel. As usual there are lots of great presentations (besides mine) and I highly recommend it. You can get use the code SPEAKPAW010 to get a 15% discount off a two-day pass and find more details at predictiveanalyticsworld.com. If you come, come by and say hi after my presentation or while I am introducing people in Track 2 on the first day.

And if you are interested in analytics, don't forget the study I am doing with B-Eye Network on business analytics - it will discuss the motivation for adopting business analytics and how you should approach the evaluation of business analytics as well as how business analytics fit within an enterprise and business architecture, risks and issues, benefits and challenges and more. You can help by taking the survey - http://www.zoomerang.com/Survey/?p=WEB22A3HRGXRBS.

I am working with the folks at B-eye Network and sponsors Oracle, SAS, Aha!, Adaptive and Fuzzy Logix on some research - Business Analytics: Putting Analytics To Work.There is growing interest in the power of analytics, especially predictive analytics, to improve business operations. The use of data mining and analytic techniques in operational systems is moving beyond its early adopter base in financial services and into the mainstream. As companies adopt business analytic techniques they struggle with the balance between using these techniques to improve reporting and dashboards ("Predictive Reporting" as it is sometimes called) and using them to improve systems and thus every individual transaction ("Business Analytics" or "Decision Management"). A clear understanding of what business analytics are, how to use them, and the compelling business value of doing so is called for. Hence the research.

The study will describe business analytics and what should you expect from a business analytics vendor. It will discuss the motivation for adopting business analytics and how you should approach the evaluation of business analytics as well as how business analytics fit within an enterprise and business architecture. It will discuss risks and issues and describe the benefits and challenges based on real customer experience. Finally it will discuss the kinds of decisions that will show a positive return on business analytics and how business analytics can change businesses fundamentally.

All in all it should be a lot of fun to write and I am looking forward to completing it. In the meantime you can help by taking the survey - http://www.zoomerang.com/Survey/?p=WEB22A3HRGXRBS.

Look for the report in a couple of months on BeyeResearch.

Quick note to let you know that the early bird pricing for Predictive Analytics World expires this week so it's time to register. You can also use my discount code - SPEAKPAW010 - to get a 15% discount. I am running a workshop the day before (you should be there) and speaking/moderating on the first day.

I got a chance to catch up with Aha! recently. Aha! is based in the Denver Tech Center and was founded back in 2006. Aha!'s premise is that it is now possible to build analytics into a platform and to focus on how to operationalize predictive models and deliver analytics within business processes. Initial customers are in healthcare, telecom, travel and transportation. Their aim is to deliver a complete analytics management system. The pain points of traditional BI solutions that they address are: limited use and access (big focus on self-service), long time to value (SaaS platform or rapid start up with minimal IT), rear-view focus (predictive analytics), piles of data (model-based analytics) and Excel/scattered data (single network). They focus on being "dynamic and aligned" and focused on business users.

Their market is the $2-$3Bn "business embedded analytics" that is part of the overall $26Bn global analytics market. In particular, they provide non power analytic users with access to analytics without having to obtain specialized skills. They see themselves helping the vast majority of business users who don't use analytics today - the people that dominate the operations of a company like marketing managers, sales managers, customer care managers, product managers, marketers, engineers, and operations specialists. Financials matter to these folks but they don't dominate the way they do with "traditional" financial department analytics users.

Aha! sells direct as a SaaS offering (setup fees and subscription), offers model development and data discovery services and licenses through OEM/SI partners. Partners are typically domain experts and vertically focused.

Some example customers include: a telecom company using analytics to handle the ROI of proactively building out a fiber network and to optimize sales and marketing to light up this fiber; a telecom handling customer retention, product segmentation and customer experience satisfaction; a healthcare company working on customer retention and acquisition.

Their offering (Axel) is a SaaS multi-tenant, multi-hosted system. It is designed to bring models into the business process - business process based models - make the analytics actionable and close the loop between analytics to actions. The whole thing is based on KPIs and designed to help companies actually act on their strategy, using a KPI model that runs from head office strategy to the front line. The platform has 5 core elements:

• Language
  The Aha! Expressions analytic model definition language that allows business analysts to build the models
• Dynamic services
  Secure, multi-tenant, forecasting, simulation and optimization
• Visualization
  Self-service, near real-time and model driven
• Data Engine
  Profiler, designer, ETL, Smart Pub/Sub
• Extensions
  Support for third parties to extend and integrate the platform

The basic process looks like this (for a healthcare member retention example):

1. Customer profile, billing, survey and claims data is used to create a model data file
2. Predictive models are developed based on this data
3. Customers are scored using these models
4. Contact and campaign management define available actions based on these scores
5. KPI-based models are developed using the same data
6. Collaborative analytics link all this together to support decision making and drive ROI

The target for this customer was to reduce churn. They were up and running in 60 days, improved retention by 7.5% (v target of 3%), improved new member retention by 9%. NPV of $43M in a single enrollment period and an all-in ROI of 2447%. This was recognized at the World Health Congress as a top example of using predictive analytics to drive member retention and satisfaction. Users ranged from call center operations to VP level executives.

The model data was used to create retention or churn scores for each customer that were loaded into the operational system in batch. These scores can be updated regularly from the model data file and can be calculated live based on intra-day data or, in theory, even during a conversation (using a standard web-services interface). The use of this model is much the same as the use of any other predictive model except that the data is tightly coupled with the KPI hierarchy. Models can be built from and evaluated against the historical data that drives the KPIs, so that users start off with a valid historical base. Axel also provides a stochastic enrichment engine ( Monte Carlo simulation with category selection, probability, and triangular distributions) that supports PMML, allowing models built outside to be imported using PMML. Models can also be generated via an Microsoft Excel Template.

Aha! is driven by a KPI model hierarchy. In the case of this healthcare company it was Retention Campaign (Strategic), then the health plan a member was in (Tactical) then events within a member lifecycle (Operational). This drives how the data is viewed and KPIs - in this case customer retention measures of various kinds - are tracked against this hierarchy. So, for instance, each KPI could be viewed with respect to a specific member lifecycle step, a particular plan or a particular campaign.

Each KPI has a calculation defined for it and are calculated dynamically. In addition to mathematical calculations, the Expressions language also provides addition functionality that supports the calculation of KPIs based on Year to Date, Quarter to Date, Month to Date, Sum of values for a defined period, Average of values for a defined period, etc.

The interface allows different reference periods to be selected and the KPIs to be viewed within that period along with measures like averages, high/low values for the period, goals etc. For instance, this customer saw a lot of new members were signing up but then being lost. The prediction showed that the trend would clearly exceed their target for such losses and allowed them to see the impact on all their KPIs. This provoked a focus on the reasons for this and they found an external verification service that was needlessly disqualifying people. They had no expectation that this would be a problem and the tool allowed them both to spot it and see the impact on their KPIs quickly enough to take action before the open enrollment period was completed and the opportunity to fix it lost.

The most interesting thing about Aha! for me is the tie to a formal model of KPIs that drive from a high level to an operational level. This allows impact analysis and decision making to be clearly linked to the objectives set at different levels.

For more information on Aha!, you can visit their website at www.ahasoftware.com or download their paper on Business Embedded Analytics.

Aha! is one of the sponsors of some research I am conducting with B-Eye Network and you can participate by taking the survey athttp://www.zoomerang.com/Survey/?p=WEB22A3HRGXRBS. You can find more reviews of products on my blog at http://jtonedm.com/category/product-news/

One of my google alerts pointed me to this thread on Oracle's discussion forums OTN Discussion Forums : Text Mining Dictionary .... I got the alert because one of friends at Oracle responded to the question "Where to Start?" by quoting the book (Smart (Enough) Systems) I wrote with Neil Raden:

Wrong: Catalog everything you have, and decide what data is important.
Right: Work backward from the solution, define the problem explicitly, and map out the data needed to populate the investigation and models.

This was one of Neil's bon mots and I was glad to be reminded of it. With analytics - executable analytics, business analytics, predictive analytics or any other kind of analytics - begin with the decision in mind. Figure out what it is you are trying to do, which decision you are trying to improve and work into your analytics and data from there. Be driven by your business needs, not by the data you have. You may find that you don't need to integrate this data source or clean that one to improve the decisions that drive your business. You may find that you don't even own the data you need and will need to go shopping for it. But if you don't start with the end in mind, you will never know.

James Kobielus had a nice list of Advanced Analytics Predictions For 2010 over on the Forrester blog. As usual James is thought provoking with some interesting predictions. Let's start with the one's with which I agree most strongly.

• Advanced analytics sinks deep roots in the data warehouse
  Absolutely. In database/in warehouse analytics will become more and more important and the analytical processing of streaming data likewise. However the way data is stored in warehouses will have to change too, not just the way analytics are done. Too many warehouses and marts today store summary data, rollups or data where the crucial time dimension is obscured. No matter how powerful the analytic engines get, this will have to change and warehouses will have to store the low-level transactional data that analytics need.
  
• User-friendly predictive modeling comes to the information workplace
  Yup. While I think there will continue to be a role for experts in building models and that executing predictive models in operational systems is at least equally important, knowledge workers are going to expect tools that let them build predictive analytics for themselves.
  
• Social network analysis bring powerful predictive analysis to the online economy
  Yes but not only t the online economy. Social network analysis is a powerful tool in telcos (see this piece on using call detail records to develop networks) and fraud detection already. Social network analysis does not require Social Networks!
  
• Low-cost data warehousing delivers fast analytics to the midmarket
  Maybe. Bringing analytics to the midmarket will be more about packaging the analytics up and making them easy to consume than about appliances.
• Self-service operational BI puts information workers in driver's seat
  I don't think this one is that compelling and I don't see most users demanding these tools. Putting information workers in the driver's seat requires making the BI tools vanish into the day to day systems and processes, not just making them self-service. Most business people want to do what they always want to do which is run their business more effectively. If tools can help with that then they will use them, otherwise not. The number who want to build mashups or self-serve on BI is a small and fairly geeky subset in my experience. Personally I don't expect the major BI vendors to make anything like enough progress in making their tools "vanish" into the systems business users use every day to deliver on this one.
  

And I know he had one more (Data warehousing virtualizing into the cloud)  but I don't have an opinion about that one.

Jeff Jonas had a great post on his blog recently, Your movements speak for themselves: Space-Time Travel Data is Analytic Super-Food! in which he made the point that:

Mobile devices in America are generating something like 600 billion geo-spatially tagged transactions per day.

With such huge volumes involved, this information is only going to be useful if it can be analyzed and used. It's real time, high volume and constantly changing. Displaying it on a dashboard or putting it in a report is not going to add much value. Use it to drive real-time decisioning, though, and you could start to add some real value. Of course this leads to concerns of surveillance and I, like Jeff, believe:

Such a surveillance intensive future is inevitable, irreversible and as I have said before here ... irresistible.

But if your customers are generating, or could be generating this kind of information, what could you do with it? Well you could use it to recommend places to go, you could use it to target offers and advertising, you could use it to schedule and route deliveries or repair crews when they are needed, or... well, lots of things.

The point is that your customers ARE generating this kind of information and the question is what are you going to do about it?

My friends at Zementis have just launched support for executing predictive analytic models in Excel - check out Predictive Analytics at your fingertips: Scoring data in Microsoft Office Excel. While not, exactly, a high-volume transaction environment, Excel is an interesting place for executing predictive models and I like the way the folks at Zementis have done it. The integration of their standard deployment engine means that IT departments have some options - cloud or on premise - for running the scoring engine while still pushing predictive analytics into Excel. I am increasingly convinced that delivering the same rules and analytics to decision support, through Excel say, as are used in automated decisioning systems is important. Most decisioning systems don't handle 100% of transactions so people will be handling the exceptions and it will be useful to them to have access to the same decisioning infrastructure.

I hosted a panel last week on predictive analytics at the Business Analytics Summit. I was joined by Richard Boire of the Boire-Filler Group, Jean-Paul Isson of Monster.com and Michael Berry of Data Miners (and author of Data Mining Techniques, one of my favorite Data mining books). I asked a series of questions and we got some great answers from the experts:

• How is your organization using predictive analytics and what has been the business value of doing so?
  • Richard's customers use predictive analytics for acquisition models that target not only high responding prospects but also prospects that will be of high value to that organization once they become customers, retention Models that target high value, high risk customers and upsell Models that allow organizations to target customers most likely to become higher-value type customers among others.
  • Monster.com uses predictive analytics for customer intimacy, customer satisfaction, customer retention, customer up sell and wallet share growth, customer acquisition, pricing, sales coverage optimization and product development.
  • Michael focused on some more general points. He made the great point, if basic, that all predictive analytics focus on the future because that's the only place you can have an effect and also pointed out that the business definition is critical. For instance, predicting which acquisition channel has the highest trial subscription sign up likelihood is potentially much less useful that predicting which channel is most likely to acquire customers that will keep a subscription beyond a trial period.
• What are the challenges you have faced implementing Predictive analytics in your business?
  • Michael emphasized cultural and educational challenges - that this is a new way of doing things and companies often resist things that are not "our way". The inability to find appropriate data in the right format was another big issue.
  • Richard talked about obtaining buy-in and engagement from key stakeholders, the challenges of data and the value of having the right team to effectively implement predictive analytics. The absence of numeracy, of basic understanding of the power and limitations of the models was another big challenge.
  • Jean-Paul also emphasized data quality and availability, especially because different countries and systems define things differently. A lack for application systems integration and standardization and of effective change management across regions can also be a problem, though the recent recession has helped with the second by making people more receptive to anything that might help.
• How did you sell predictive analytics - how do you demonstrate the value of predictive analytics to the various stakeholders within your business?
  • Richard suggested conducting sensitivity and business analysis to demonstrate monetary potential of project as well as identifying stakeholders who are engaged with the data and working with them to prove your case. A project that is a quick win in terms of ROI and implementation also really helps.
  • Jean-Paul emphasized taking baby steps - starting with the basics and always have something meaningful to deliver. Showing the ROI of a model on a small group of customers (a smaller country or region for instance) also really helped.
  • Michael said to focus on showing how the model will help them do what they do and made the point that he often finds he is the first to look at the data, putting him solidly into discovery mode. Like Richard and Jean-Paul he emphasized the importance of linking everything to real monetary measures.
• With predictive analytics being such a hot topic, what do you think holds companies back from embracing and exploiting these techniques?
  • Richard felt that a lack of knowledge combined with a discomfort around mathematics and numbers was a big problem. Change management and adopting a new approach also cause problems.
  • Michael emphasized a lack of executive support and the need to get enough support to overcome organizational inertia. He also had a great example where existing measures can make adopting a model hard because the model will drive better overall results while driving a critical measure in the "wrong" direction.
  • Jean-Paul talked about the lack of understanding/knowledge of the real value of predictive analytics also. The attitude of old school management that they are already successful so why do they need to change and spend more money. Bad experience with IT solutions over the years and the communication skills of those proposing the idea sometimes don't help either.
• What skills set are required to achieve success with predictive analytics?
  • Richard emphasized importance of learning about the business domain, both so that effective models can be developed and so that the models can be related to measures that matter to business executives. Obviously strong quantitative/mathematical background and an ability to work easily with numbers as well as good communication and interpersonal skills were also needed.
  • Jean-Paul said that a wide variety of skills are required with programmers, statisticians and data miners, business analysts, and web developers needed to deliver the solution to end users.
  • Michael pointed out that intuition and creativity - an ability to see what's important - is necessary also.
• We wrapped up with the question what does it take to operationalize predictive analytics, to integrate predictive analytics as a regular business discipline? What are the pitfalls?
  • Richard talked about discipline, repeatability, as well as tracking and performance management. A ruthless focus on the business implementation model is also key.
  • Michael reminded us that actionability is critical - if we cannot act and act effectively on a prediction then it does us little good.
  • Jean-Paul said it takes a clear vision, human capital, collaboration, people/process/technology and a focus on the customer/user experience.

I really enjoyed the panel and I hope I have captured its essence here. If predictive analytics interests you, and it probably should, check out this white paper I wrote on Putting Predictive Analytics to Work and this webinar I recorded with Eric Seigel on Optimizing Business Decisions with Predictive Analytics. Cross-posted to BeyeNetwork and ebizQ, both of whom were media sponsors of the event.

I am hosting a panel on Predictive Analytics at the Business Analytics Summit and I got a chance to attend a session beforehand where Dave Stodder presented on performance management and Key Performance Indicators.

Dave began by emphasizing that performance management is both a business and IT issue and that it needs to link people, process and technology. Performance management is focused on how we are doing, why are these things happening and what should we be doing. KPIs are designed to track and measure within a performance management framework. At the end of the day it comes down to Drucker's comment that you cannot manage what you don't measure. As a result Performance management is driven by everything from Balanced Scorecards (Kaplan and Norton), Six Sigma, TQM and more.

Balanced KPIs keep people focused on what they should be doing, what they can do with their information, as well as providing balance between conflicting goals. They should also be based on multiple measures not just financial ones. Performance Management joins Process Management and Decision Management as a "higher power", one of the levers of improved enterprise performance. Performance management also helps bring BI from departmental usage, focused on reporting, to enterprise-driven metrics and best practices.

Three business imperatives drive demand for BI, Performance Management and KPIs:

• Establish value - find, increase and retain value
• Manage risk - identify, predict and protect
• Rebuild worker productivity - measure, manage and enhance by focusing workers on core issues

Performance Management has some overlap with operational BI. The move to an operational focus, from traditional BI to more operational BI, is often focused on improving efficiency and customer service. Getting rapid access to accurate information improves efficiency and improves customer service. This pushes more BI functionality out to front-line employees, business processes. The use of performance management and a focus on metrics can help focus this and simplify it by allowing IT to deliver simple metrics rather than complex reports.

At the same time the focus is moving to centrally managed approaches, where departmental systems were more common in the past. The move to operational BI and metric-driven performance management is driving centralization. This delivers consistency, cheaper/faster integration and makes it easier to implement data mining and predictive analytics.

Improving customer service, focusing on customer metrics, is the most important objective for 60% of folks surveyed by Ventana Research. This focus on customers means more sources (because customer data is scattered) and on more real-time data (because customers keep doing things). Interestingly customer contact centers are emblematic of the challenges with KPIs. For instance, though the agents are measured and managed very tightly, their supervisors are not. Supervisors don't feel they can impact customer service directly so they don't see the metrics as relevant to them. KPIs must be designed to match what the person can impact (and, I would add, give them the ability to change the systems that affect the metric when this is necessary).

Accountability is critical for KPIs. Is it reasonable to hold someone accountable to a particular metric? Are the people who come up with the metrics held accountable for their implementation? Who needs to see what - are you keeping people focused on metrics that fall into their area of responsibility?

Proliferation is a second issue. Just like reports, too many KPIs can be distracting rather than useful. Vague implementation without real accountability and control will result in metrics that are just noise.

A central focus on well defined metrics can also really help after mergers and acquisitions - it can be easier to develop integrated metrics than integrated reporting. A focus on metrics also tends to deliver a top-down view not the bottom-up view typical of reporting, and improve alignment.

Human psychology is critical in performance management. Will people focus only on achieving the goal or will they be more thoughtful about what is really needed? Will people understand why those metrics matter? Do people want their performance to be transparent and at what level?

Keep the number of KPIs reasonable, make sure people understand what is driving the KPIs and think continuous change.

Cross-posted to ebizQ and BeyeNetwork as both were media sponsors of the Business Analytics Summit