Adam Gartenberg's Blog

Business Analytics and Optimization, IBM and Social Marketing

IOD 2013 Day 1 General Session

As I've done the past few years, below are my notes from the IOD 2013 opening general session.  Please see my other posts for more details and links on today's announcements.

Jake Porway
What does big data look like?  Sometimes it gets described as spreadsheet, or maybe the ubiquitous tunnel of 1s and 0s.  Instead, he prefers to think of it this way.  Go back to 2000 when renting movies meant going to a retail rental store, with no idea if you're going to find the movie that want, with no ability to see ratings or reviews, and so on.  And compare that to our options today.

Another example he gave of big data in action is a chart comparing the prevalence of the flu as reported by people mentioning flu-like symptoms on Twitter vs. the much more extensive (and expensive) tracking done by the CDC.  The graphs follow very similar patterns... with Twitter often leading the CDC results by about two weeks.  (Note:  It looks like this is the research he referenced:  You Are What You Tweet: Analyzing Twitter for Public Health, by two researchers from Johns Hopkins University.  And while Twitter likely won't have the robustness needed to replace something like the CDC tracking, it certainly is an interesting comparison to make.)

Image:IOD 2013 Day 1 General Session
Moving on to other kind of possibilities offered up by Big Data, Jake mentioned a recent hackathon that produced some interesting solutions, but for the most part focused on solving problems that are important to data scientists (e.g., finding a better parking spot).  Now, what can happen when you combine that capability with people who are doing things like finding water in developing countries.  There's a lot of data out there - how do we put it to use?

DataKind is a non-profit connecting pro bono data scientists with social organizations.  Things that might be easy for people in the room we are in can be transformative for organizations that don't have those skills.  For example:

  • DC Action for Kids.  They have a lot of data on child well-being, and teamed up with data scientists from the Washington Post and World Economic Forum.  Together, they turned largely incomprehensible rows of data into an interactive map of child well being across the city.
  • NYC Parks - Wanted to know if the things they're doing (e.g., cutting down tree limbs) actually make a difference (e.g., preventing future problems).  They were matched up with a data scientist from a NY ad agency, and found that there were 32% fewer emergencies reported in areas where they had pruned trees compared to those where they hadn't..
  • The World Bank - teamed up with data scientists to look at food price data.  They scraped supermarket website and mobile shopping apps to track prices for a basic 2000 calorie diet across Africa.  They also looked at rice prices.  Now they're looking to see if the amount of light visible from satellites at night can help them track progress against poverty..

Jake concluded by pointing out that at this point, we're just limited by our creativity.  The data is there, technology is there.  How are you going to do it to transform your business, your industry, your world?

Robert LeBlanc, SVP, Middleware Software

35% of those in auditorium are here for the first time.

When look at technology transforming businesses:
  • Enterprise Mobility
  • Cloud Computing
  • Big Data and Analytics
  • Social Business

These are all enablers:
  • Enabling us to be much more client centric
  • Insight-driven - Embracing the data, don't fight the data.  
  • Forge deeper relationships - just like in personal lives, gives us better insights into individuals and customers
  • Constantly Adapt - We're still at the cusp, only seeing the tip of the iceberg.  We need to help our businesses change.

There are a lot of untapped resources out there.  Data for data's sake - that's interesting.  But if we can analyze and get insight, we can really change things.

It used to be I was valuable for what I knew.  Now I'm valuable for what I share.  The end goal is increasing value for clients and employees.

When think big data, don't just think volume.  Think about variety - unstructured data, and especially the impact of social data.  Think about velocity - real-time data.  You want to be forward looking, not driving but what see in rear-view mirror.  You want to be able to react to what your clients want.  Think about veracity - think about the quality of all the data that's out there.  Make sure you're making decisions with the knowledge of what that data is and where it came from.

Big Data and Analytics is enabling the next generation of systems.  "Systems of Engagement" open up a new set of interactions.  It allows you to add value at every single point of that interaction, whether with employees, clients, potential clients.  When you bring that together with Systems of Record you get a continuum of Systems of Interaction.

Why act now?  According to a recent study:
  • 75% of clients cite growth as the key source of value from analytics
  • 46% of respondents were impact by a cyber security breach over the past 24 months
  • Only 1 in 5 organizations allocate more than 50% of IT budgets to new projects

From Nationwide Mutual Insurance Company- VP of customer analytics, Wes Hunt
  • Nationwide sells full range of auto, home, commercial insurance, life and retirement
  • How deliver promise of "on your side" to consumers?
  • It means:  We know you, we care about you, we're easy to do business with.  These are consistent across categories and all members.  Strive to know you as if you are only member of the company.
  • Made decision in 2008 to deliver on seamless customer experience across all channels and touch points.
  • Where did Nationwide start?
  • Started with Info Mgmt foundation, then realized that to deliver on the promise, every front line representative needed history on relationship.  Had 27 different front-end systems.  When integrated this info, though, unlocked power of amazing insights.
  • Today, have learned that big data without analytics is an extraordinary tax on organization, and analytics without big data will likely lead to wrong conclusion.  Made realization that are better together.
  • Why did Nationwide pick IBM?
  • Use IBM MDM platform to house customer info. SPSS for predictive modeling. IBM consulting to help design, build and implement the solution.  Did with great partnership.  Can't tell where IBM starts and Nationwide ends.
  • Is there payback?
  • Focused analytic agenda on understanding and instrumenting critical moments of truth in customer buying journey.  Today, have 360 degree view of customer information with deep insights based on behaviors and actions that improve performance.  Integrated across internet, call center and local agents.  Rated #1 in customer experience in recent study.  Have seen substantial gain in what see as most important measure - customer satisfaction.
  • What Nationwide learned:
  1. Build strong data foundation, one that will handle change.  5 years ago didn't know every problem that would need to solve.
  2. Grow your talent.  We don't have enough talent to solve all the problems, and people are the ones who yield results.
  3. Big data and Analytics is not a corporate strategy, but rather is an enabler of our strategy.  Find the cross-hairs and stand in the middle.
  4. Secret ingredient for success is trust.  Trust takes 3 forms:  Trust by end users in the insight.  Trust by analysts in the data.  Trust between business and IT.  Trust is the bridge between insight and action.

How do pharma reps determine which doctors to visit?  Historically, visit doctors that already prescribe their medications.
Next best action, with mobile data.  Combined doctors with high propensity to prescribe more medication with geographic location.
400% ROI in first year

GM Business Analytics - Les Rechan
GM Information Management - Bob Picciano

Here to talk about our performance - continuously transform our organizations.
Over past few months, spent time with over 1,000 clients to find success factors for analytics.

1.  Infuse culture of analytics everywhere in the organization - Imagine It
2.  Platform for big data and analytics - Realize It
3. Security, privacy, risk - Trust it

Imagine it - infusing analytics everywhere
  • Will need 20% increase in skills.  Only 1 in 10 organization have skills they need
  • Roles - City of NY have announced chief analytics officer.
  • Mastery - Analytics quotient of departments and enterprises

Realize it - Curate all big data and prepare it to be analyzed
  • Need new architecture for the 4 v's.
  • IBM has combined that entire big data and analytics platform in one big vision.  
  • Data in motion is like being able to stop time on transactional information.  That's what infosphere streams does - stops time for data in motion to apply analytics to that data and then move forward

What we can deliver:
  • Automated stack report that looks at speed and capability of various infrastructures.  4x performance of anyone else's Hadoop environment out of the box
  • DB2 BLU + Power + Cognos BI.  38x average acceleration of database quires for reporting

Need platform that means you can trust the data, that you can ensure privacy and security of everything you're doing.  Make risk aware decisions - model exposure and understand variability.  

Capabilities that delivered throughout course of the year.
April - Announced BLU acceleration capability.  Breakthroughs in columnar processing.
May - Great advances with Watson.  
September - Governance, visualization of
Other acquisitions Star Analytics, Daeja, The Now Factory

IBM Cloud now hosts more top websites than anyone else in the world.  Cloud now gives customers more ways to customize their websites than anyone else.

Announcements at IOD
  • InfoSphere Data Privacy for Hadoop platform - data privacy, masking, protection for Hadoop.
  • Petabyte Scale Exploration for InfoSphere Data Explorer
  • Write Once, Run Anywhere SQL
  • Analytics for System z

IBM Optimizes Infrastructure for Consumability
  • Fast on Fast Analytics - BLU Acceleration - Power Systems Edition
  • Accelerated Deployment - BI Pattern with BLU Acceleration SPSS Modeler
  • IT Operations Analytics - SmartCloud Analytics Predictive Insights
  • Security Intelligence - Event Management early access Program

Redefines Analytics for the Cloud
  • Agile Warehousing - BLU Acceleration for Cloud Early Access Preview
  • Performance Mgmt and Risk Analytics
  • Next Generation of Data Disocverty
  • Content Manager

Project Neo - next generation knowledge discovery - natural language, guide the user

Helping with skills:  
  • Research facilities so you can bring your data in and use best minds in data scientists to gain insights.  Have opened 9 analytics centers around the world to let clients come in, leverage resources, do proof of concept work.  Over 9,000 consultants with over 30,000 engagements in 170 countries around the world.
  • Over 1000 univ partnerships and new talent initiative.  Over 2,500 business partners.  Over 100,000 people have accessed analytics zone

VP of Big Data Integration and Governance, Inhi Cho Suh
US Tennis Open is the largest annual sporting event in world.  700,000 people in NYC, 200 countries broadcast around the world.

What was USTA hoping to achieve?
  • They asked us to help us create world class experience as dynamic as live event, across all digital platforms
  • Mission of USTA is to grow the number of people watching and participating in the sport of tennis.  Offer a chance for fans and players to share love of the game.
  • For the fans:
  • 2 million unique visitors visited the site (most during the 2 weeks of the event).  53 million visits, 419 million page views.  
  • Hottest new applications is called slam tracker.  Allows fan to be in the driver seat - play with match data along with prior data.
  • Leverage SPSS to mine 41 million data points from 8 years of grand slam data.  Determine 3 keys for each player before going to upcoming match.
  • They would only release it once players got ready to go on court so as not to influence the players
  • Operations behind digital experience.  Brian O'Connell - Senior Tech Staff Member, IBM Continuous Availability Services
  • Looking for ways to reduce operating costs.  Previous methods were spreadsheets with year to year growth, and not keeping up with growth in mobile and social.  
  • How handle variability?
  • Embraced uncertainty.  Every day in live sports is different than the day before.  Predictive cloud continuously forecasts needs and adjusts provisioning using analytics.  
  • Example - looking at snapshot in time for this year's US Open.  Don't actually know the times of play.  Forecast based on social input, weather, etc.  Use InfoSphere Streams to ingest social conversation around US Open to find instantaneous changes in player popularity.  
  • Also using IBM BigInsights to look at text analytics across entire website.  Which players are being talked about most (or even weather delay).  Will tell how popular each player is and therefore might say how much provisioning would need when that player plays.
  • More than 95% accurate
  • Application in other industries
  • Retial, insurance... even financial arbitrage

Next turn to different industry - healthcare panel:
Dr. Timothy Buchman - Emory Critical Care Center
Matt Muhart - EVP, CFO Memorial Healthcare System

Buchman - Pressured to deliver better, safer care at lower costs.  need to find a way to do that
Muhart - One of largest public healthcare system in country.  Significant downward pressure on revenue streams and demand from public to see more transparence on pricing and outcomes

Muhart - Suspected fraud among vendors, hired investigators and found individuals that had created false front companies and were rigging the bidding.  After confirmed suspicions, brought in FBI and IRS.  Netted 12 people who pled or were found guilty.
Started first foray into big data.  Was manual.  Had a firm come in and evaluate vendors - who they were, if they were related to any employees, if they were related to each other, any criminal background.  Manual process was overwhelming and too costly.  Team was gathering binders of data and realized that likely had red flags that missed.
Reached out to IT consulting firm, and asked them to build system for them.  System searches 800 diff. databases, employee database, vendor db, and flags anything that could be possible conflict of interest.  have precluded a number of vendors for a variety of reasons.
ROI?  Bad guys required to pay $2.6M in restitution, and actual fraud likely exceeded that.  The risk to reputation is significant, too, as a public health organization.

Buchman - Go to a bedside and see a patient in critical care bed, have 6 different IV pumps, streams of data coming off all of the sensors.  No person can possibly make sense of that.  Looking at how can comprehend all that data and forecast it.

Why IBM?
Muhart - Underlying tool was underlying development with law enforcement
Buchman - three reasons:
  • Research infrastructure of IBM
  • Partnering not just with us, but with third party on how to pull data in
  • Analytics itself - no other engine out there like InfoSphere Streams to grab data in motion, analyze it, and deliver to care givers who need it.

What do you envision next?
Muhart - Going from compensation based on what do to compensation to how healthy community is.  Focusing on how to identify where issues are and fix them to keep people from coming in
Buchman - Imagine if knew what most valued based drug was to deliver to that specific patient.  Imagine if knew how long that patient would need to be in unit.  Imagine if knew what device they would need and get it there just when need it.  And if all that goes into database so that can deliver better care tomorrow.

What advice have for audience?
Muhart - Data is a very lazy asset..  It needs to be taken to gym, exercised to be sure you get value from it.
Buchman - Get to the data, know it's out there, go to silos in org and imagine how might putt i together in new ways.

Jake Porway
If you're sitting in the audience and wondering "Do I really need to take advantage of Big Data and Analytics" is kind of like sitting in the '90s saying "Do I really need to take advantage of computers?"