Adam Gartenberg's Blog

Business Analytics and Optimization, IBM and Social Marketing

Live Blogging IOD2009 Day 2 General Session


The Day 2 opening session is getting started.  

Mary Garrett - Chief Marketing Officer, IBM Sales and Distribution is hosting today.

In Singapore, 1500 cameras monitor the traffic.  Tow trucks arrive at accidents within 15 minutes, and you can predict when the bus will show up to within 1 minute.  Singapore has smart transportation.  Copenhagen has Smart Utility Network that reduces power outages to 5 minutes.  In London, 45M gallons of water a day are lost to leakage.  They were able to reduce it by 20% by putting smart meters in people's homes.

Steve Mills coming to the stage.

This morning want to make connection with you on idea of information-led transformation and smarter planet.  As 36 year veteran of IT industry and lived through many eras of computers.  Like expression "information-led transformation"  Have seen flow as move from era of highly centralized computing, personal computers, client/server, world-wide web.  This decade's incredible growth and expansion around business process integration and process-led transformation.  On cusp of another change, and that will be information-led.  Doesn't deny what done before, but points to where can increased value.

Smarter planet - taking advantage of art of possible about what info technology makes possible today.  Economies coming down to where can take advantage of all this technology, embed it, make it all more intelligent.  The time is right.  Idea of smarter planet - felt it would resonate with people that there is something we can do to fundamentally change and improve the world we live in.

Information is at core of what talking about.  New Intelligence - how apply it to make world greener, working smarter.

A smart planet is one that is instrumented, interconnected and intelligent.

What does it mean to be smarter?  

Begins with ideas of instrumentation - sensors and metering.  If want control of something, must have visibility.  Otherwise can't tell what's going right and what's going wrong, and price points are coming down to point where affordable.  And maintenance benefits often will more than pay off investment - examples will give all have payoff in less than 6 months payback in preventive maintenance savings alone, before system was actually optimized.

Data integration, filter out the noise.

Comparison of historical data, with newly collected data.  Data modeling and analytics - to find clusters of data, derive insight from it.  Then get to and present to people that need to take action on it.  Of course, feedback loop as go back with it.

Examples:
AIrbus - big aircraft manufacturer.  Have massive problem with literally millions of parts that go into aircraft that have, different suppliers.  Very assembly driven, with thousands of sub-suppliers and assemblers.  Huge cost tied up in process and inventory.  If missing critical part, whole process that relies on just in time delivery breaks down.

Full RFID sensing environment where all material is tagged through subassembly supplier processes, moved to right locations, staged into manufacturing plant, and utilized to make sure have just in time parts availability to build planes they need to build.  Ties into ERP systems.  Terabytes and Terabytes of data, and gives insight into efficiency of supply chain so can systematically over time bring down cost of building airplanes.

Another example - water management.  Availability of clean water supplies.

Galway Bay, Ireland has a project under way to monitor through sensors and cameras the health of the bay, condition of fresh-water runoff into the bay, and making sure the clean water requirements are being complied to.  Understanding wave action so can tap into ocean wave activity for energy generation in the future.

Another example - food processing and delivery.  HRAFN is major food processor - meat processing business, from farm to table.  Concerned about freshness and consistency of product delivering, with of course implications on health.  Literally attacking on ranch, with RFID tagging of animals, right out through processding and out to the retailer.  Significant RFID project, huge amounts of data, lot of analytics behind it.  Great success story, and fascinating as think about RFID - beyond what see at a place at Airbus.  Literally tagging livestock.

Announcing that HRAFN is the recipient of the 2009 Outstanding Smarter Planet Solution award.

Building smarter cities.

Smarter Government - Alameda County SOcial Services Agency.  Have one of most forward-looking set of services - family welfare, child welfare services.  Pulling together county data from six disparate state and county systems integrated into a data warehouse.  The end result is real-time info for health care workers, and ability to reduced time spent by investigators for one case from 5 months to one minute.  Also award winner for Outstanding Smarter Planet Solution.

Smarter Transportation.  Are smarter transportation projects taking place around the world.  Reducing the amount of carbon emissions, increased use of public transportation.  

Smarter Public Safety - New York Police Department - 150 data bases, petabytes of data analyzed and mined to prevent crimes.  

Geisinger Health System - Self-contained health system in Pennsylvania.  What are treatments, outcomes, all organized and made consistent so physicians can look at your personal data as well as outcome data for people with similar symptoms or conditions or medical history.

Hamilton County Dept of Education in Chattanooga, TN - Using SPSS data mining software to take a deeper look at studen tpefrmance by combining data sources and exploring variables beyond what state reports provided.  Able to identify warning signs of low performance and proactively intervene, seeing a 11.5% reduction in high school dropouts.

DONG Energy in Denmark - Lesson outage times by 25-50% and seeing up to 90% capital savings.

Information will be the transforming technology for years to come.  Infinitely flexible, and can make world's systems as efficient as back office systems and processes have been automating up to now.

Customer panel moderated by Mary Garett:  George - Cardinal Health, Svien Sagatun - StatoilHydro, Dr. David Acheson - Leavitt Partners (formerly head of FDA Food Safety division)

Q:  Share with us the things you've seen and how information is used in food area
A:  When think about food safety and sustainability, need to recognize this is one of most complex systems need to dela with.  Domestic food supply in US.  150,000 manufacturers of food - excluding farms, retailers, restaurants.  Now importing 15% of food consume in the US; Agriculture  - 80%, produce - 40-50%.  200,000 foreign distributors and manufacturers (again, excluding farms).  Such a dramatically enormous system designed to put food on the table 3 times a day, and we expect it to be 100% safe.  Knowing that supply chain from farm to fork.

Q:  Any experience in how secure that supply chain?
A:  Yes, important - need to be safe from natural and unnatural events.
A:  Cardinal health - look at how secure supply chain of pharmaceuticals - make sure don't have counterfeit drugs, allow to do analytics to make sure.
A:  StatoilHydro - Better to predict and act than see and respond.  Is how want to act - use investments to avoid accidents all together, rather than analyze why they happen.

Q:  How ensure safety in pharmaceuticals
A:  Criminals are getting smarter, so have loss to theft, and how track so when something is stolen, can work with authorities to track criminals stealing those products.

Q:  How take info and apply it, when look at applying business analytics
A:  Statoil - Industry is risk based industry.  If drill exploration well, might cost to $100-250 million, and frequently end up with dry well.  On other side, tremendous opportunities to optimize recovery and production with just percent or fraction of percent if can predict and analyze better.  
A: Cardinal - How apply business analytics to commitment to customers. Supply pharma at high levels of service.  Need to make sure have inventory of product at right time.  Want better analytics so can look at inventory position and ensure have product in supply chain where need.

Q:  A lot of opportunity to apply analytics in food area.
A:  One of things that keep big food processors awake at night is how safe and secure things are in their supply chain.  At other end, might be small farmer or producer - in US or outside - where records (if kept) are on pieces of paper in a shoebox.  Might not even have computer connectivity.  Examples for context - 1.  Peanut butter outbreak had in the US that sickened 700 people and killed 9.  All linked back to small peanut butter manufacturer in Georgia.  This was a small manufacturer, but was distributing throughout the US.  Going to over 400 different retailers and manufacturers that were turning it into other products.  Because of that, and the lack of connectivity down system, when this problem began in January, there were many people down supply chain that didn't even know they had products from that supplier.  Recall ran through June, with many people continuing to get sick.  It's a data flow.  Problems could have been prevented with preventive controls.

2006 - big spinach outbreak in the US, and spinach consumption still not back to where it was.  This occurred on small ranch in California, and went all over US.  Again - if could have preventive controls on supply chain, would have very different outcome.

All about prevention - having data and analytics to prevent it from happening, and shut it down as quickly as can when it does.  Technology comes in with having really extensible traceability system.  Steve gave a great example of that in Norway.  The technical requirements are clearly doable - they've shown that.

Cardinal Health - Just starting our journey on starting up infrastructure laboratory on traceability.

Q:  Talked about information and predictability.  As talk about solutions all working on, Steve mentioned instrumented, interconnected, intelligent.  How have you been looking at these solutions within your enterprises.  Peoplea re at different stages of maturation.
A:  Cardinal Health - How build out fabric of solutions.  Had opportunity to make investments in traceability server, master data mgmt.  Bring together to fabric of capabilities so can respond to any challenges that have.  Have signficant investment in legacy systems - how pull out that large amount of data and apply analytics to it.  Where next opportunities for us and how use to address business challenges.
Q:  How long on this journey?
A:  Goes back 2-3 years.  See roadmpa going many years into the future.
Q:  Are business leaders on journey with you?
A:  Yes - always need to tie back to business case.  Have exec leadership to make sure can make investments to produce most cost effective means for health care

Q:  How about StatOil
A:  Do a lot of work in remote locations. Can't bring extra personnel out to those locations, so need to bring the data to the people.  Have invested heavily in instrumentation.  Hundreds and hundreds of km of fiber optics from at-sea locations.  
Q:  Investments?
A:  Investments are so big, that 0.1% savings could justify the investment
Q:  How about at the FDA?
A:  Dealing with spectacularly complex environment.  Part of challenge is 350,000 foreign and domestic manufacturers.  What doing at these facilities.  Every one has to be registered with FDA.  Part of data aspects is acquiring that registration information.  Right now it's just who are you, what are you, and very generic description of what doing.  Better than not knowing who's out there, but a lot of opportunity to improve on the situation.  FDA has 3,000 inspectors for those 350,000 facilities.  At the rate FDA was going, would take 1900 years to inspect all those facilities.  How could data acquisition and analysis make that better?  Could plug into what foreign gov'ts doing, where can trust that when they inspect those facilities, can upload that to the US FDA system.  Another way would be to use 3rd party vendors and enter that data into the system.  Can work to identify which suppliers are high risk vs. low risk, so when hit inspection point at 300 ports in US, can prioritize who to look at.
A:  Cardinal - Similar problem in pharma.  Had some challenges with FDA, and in working through that, now have project under way with IBM Watson Research to create tracking and data to enable reporting on suspicious activity.

Q:  Talked a lot about success that had and how using data, going from information to predictability and traceability.  If could sum up
A:  Cardinal - Vision and Focus.  Vision - appreciating IBM's vision in where investments being made today and infrastructure capabilities growing into solutions.  And making sure that that investment is tied into Cardinal's goals.  Focus - Getting technology leadership wtihin organization having high energy and commitment for extending capabilities,k and taking opportunity and driving through team with high energy
A:  Statoil - Continuous Improvement - That's focus.  Data avaialbility, benchmarking, ability to predict outcome and investment decisions
A:  Leavitt - Prevention, Risk, React.  The goal of anyone in food industry is to prevent things from going wrong in first place.  To do that need to understand risk, which end of supply chain is the problem in, is it inside/outside the company.  And need to react when things do happen.

Joining on stage - Bill Pulleyblank, VP Advanced Integration, GBS Business Analytics and Optimization

Do you remember the first time heard word H1N1?  One of least favorite 4-letter words.  Health care shows biggest challenges and opportunities on smarter planet.

Project did with University of Ontario Institute of Technology shows impact.  [video - I'll look to find a copy and link to it later.]  Are up to 16 streams of info all being watched by nursing staff, clinicians, residents, all on own timeframes and with different observations.  Saw opportunity to capture data and analyze it.  Addressing the challenge in neonatology is through ingegrated environment of InfoSphere Streams, DB2, and SolidDB.  Can analyze data in real time and take earlier proactive response.  The end result is being able to dect neonatal infection up to 24 hours earlier.  

How deal with streaming data as opposed to static data.  What if need to respond in real time as data passes.  Can plug in different analytics at different points, pulling in different parts of the stream.

Diagnostics - Geisinger collecting clinical data, making it available so Dr. in office can find data and do better job diagnosing disease.  But there's another source of data swamping physicians - published data.  Constant flow of new treatments, new symptoms, coming from around the world.  A system called "Watson" - what if could take all this info being published, collect, assimilate, figure out what's in it, so when physician thinks about possible cause of disease, can incorporate all that information to that.  So articles that matter most are brought to doctor's attention, rather than requiring that they have come across it and recall it.  Reduces cost, because if know what problem is, don't have to run unecessary tests to try to diagnose it.

Computational medicine - can you model the human heart - and not just any heart, but your heart, down to molecular level, so when considering treament, are talking specirically about what you need.  The challenge of modelling something as complex as heart is a huge computational requirement, but when can do that, can truly personalize medicine.

Getting ahead of influenza.  Constantly mutating.  Would it be possible to predict which mutations will come up in the following season?  Data analysis done on mutations, and were able to predict which mutations would come next.  Gives head start on treatment and creation of vaccine.  

DNA Transistor - Disease is very specific, so if you could take your DNA and discover your genome, would open new possibilities in treatment.  Problem is that costs a lot of money to sequence genome.  Dream is to be able to sequence an individual's genome for under $1000.  IBM research is undertaking this today.  (NY Times article on it.  While there are detractors who say it's not possible, we're confident that can.)

Can a computer win at Jeopardy?  Designing computer system to participate as a contestant on Jeopardy.  With each iteration of the software are getting closer and closer to where can take on Jeopardy champions.

Summing up - Information Analytics - take best of science and best of business and see how can bring this to individuals.

Mary Garrett back on stage.  Change is happening.  Can choose to lead.  

Tomorrow morning - focus on levers of change, another customer story, and special guest Malcolm Gladwell.

That's it for the Day 2 opening session.