San Francisco • Chicago • New York City
Berlin • Washington DC • London
Delivering on the promise of data science
Click here for upcoming PAW events

Full Agenda – London, UK – 30 November - 1 December, 2011

Day 1: Wednesday, 30 November, 2011


Registration & Coffee

[ Top of this page ] [ Agenda overview ]


Persuasion by the Numbers: Optimise Marketing Influence by Predicting It

Data driven marketing decisions are meant to maximise impact – right? Well, the only way to optimise marketing influence is to predict it. The analytical method to do this is called uplift modelling. This is a completely different animal from what most models predict: customer behaviour. Instead, uplift models predict the influence on customer behaviour gained by choosing one marketing action over another. The good news is case studies show ROI going where it has never gone before. The bad news? You need a control set… But you should have been using one anyway! The crazy part is that "marketing influence" can never be observed for any one customer, since it literally involves the inner workings of the customer's central nervous system. If influence can’t be observed, how can we possibly model and predict it?

Speaker: Eric Siegel, Ph.D., Programme Chair, Predictive Analytics World

[ Top of this page ] [ Agenda overview ]


Refreshment Break

Customer Retention with Churn Modelling
Case Study: Paychex Inc.
Combat Client Churn with Predictive Analytics

In economic conditions such as these, it is critical for businesses to have a stronghold on their client retention efforts. Historically, it has been shown that businesses that excel in this arena are often better positioned for long-term success and possess a competitive advantage. To optimise the value of retained customers it’s essential to understand which clients are a fit for retention campaigns so that the loss of time and resources is minimised. In this session, we will review how Paychex leveraged two existing models, Paychex Attrition Model and a custom built Lifetime Value Model, to create a Retention Tracking System (RTS). Since being deployed across the entire branch network, the Retention Tracking System has become an invaluable resource as offices nation-wide strive to meet, and exceed, their retention goals.

Speaker: Frank Fiorille, Director, Paychex Inc.

[ Top of this page ] [ Agenda overview ]

Human Resource Retention
Case Study: Hewlett Packard
Attrition Driver Analysis

Attrition is a very common issue with most organisations as the cost incurred by organisations is usually quite high. In a normal scenario, it’s hard for an organisation to understand which of its employees are most likely to attrite. But an early warning may help organisations build strategies to retain more people which will help to reduce cost and maintain continuity in their businesses. The objective of this project is to find out the drivers of attrition and use them to compute probability of attrition for each employee of an organisation or part of the organisation using a predictive model.

Speaker: Anindya Dey, Analytics Consultant, Hewlett Packard

Speaker: Gitali Halder, Manager - Analytics, Hewlett Packard

[ Top of this page ] [ Agenda overview ]


Platinum Sponsor Presentation
Healthcare Analytics: Predicting Problems Ahead

The National Health Service (NHS) workforce is one of the largest and most complex in the world. The staff working for the NHS are also the single most important factor in ensuring quality and safe healthcare services in the UK. The NHS is increasingly looking to data and analytic techniques to understand the relationships in this complex data landscape, and use this information to forecast future issues and support decision making. Chris is leading a project on behalf of the NHS to develop an analytic platform to support the effective delivery of services through workforce assurance.

Speaker: Chris Stirling, Partner, Deloitte

[ Top of this page ] [ Agenda overview ]

Special Featured Session
Case Study: ING Belgium
Successes and lessons in a five-year growth path in Predictive Analytics

This session will cover ING Belgium's journey towards becoming a convinced user of Predictive Analytics. While the first successes were well received by a small group of early adopters, they remained sporadic, and the impact of analytics on ING's communication policy was limited. This presentation will illustrate which major - organisational and analytical - developments influenced the usage of Predictive Analytics throughout the organisation, and eventually the success of this business unit. As a result, nowadays, predictive models are the main engine of automated campaigns towards a large variety of target audiences and channels. We present the evolution from initial model developments to industrialisation of model development and model scoring, automated monitoring and finally optimisation. From a non-technical point of view, we emphasise our hits-and-misses in achieving involvement, buy in, and usage of a battery of predictive models. We conclude by presenting some challenges that still remain.

Speaker: Geert Verstraeten, Partner, Python Predictions

Speaker: Pieter Dyserinck, Head of Business Analysts, ING Belgium

[ Top of this page ] [ Agenda overview ]


Room: Expo Hall

Track 1: Customer Retention with Churn Modelling
Case Study: IPC Media
Proving the value of Prediction in Magazine Subscriptions

In this presentation Claire and John will give a practical view of how IPC Media designed, built, tested and deployed several models which led to increased retention and profitability. They will also discuss how the diagnostic detail of the models have provided valuable customer intelligence for use in long-term IPC retention strategies.

Speaker: Claire Aspinall, Renewals Manager, IPC Media

Speaker: John McConnell, Director, Analytical People

[ Top of this page ] [ Agenda overview ]

Track 2: Financial Services
Case Study: The Royal Bank of Scotland
Banking Value From The Long Tail

While bankers pay a lot of attention to negotiating the best value from large deals at the top end of the corporate market, it is neither economical nor practical to apply this diligence to the mid-market segment given the large transactional and customer volumes. The Perito rule (80/20) applies everywhere. This results in inferior commercial disciplines and the weakening of the proposition across a large swathe of mid-market banking portfolio.

The "old school" of banking was a simple revenue and volume driven game. New regulatory challenges from ILAA and B3 have increased the divergence between income and value - particularly across customer segments. This suddenly makes the hitherto under-invested areas of banking (such as mid corporate market) very attractive. However, if extracting income from a large volume portfolio was previously difficult, it is now even more difficult to understand where the value pools lie and how to extract them!

This is where predictive analytics comes to the rescue. Gaurav will talk about how by synthesising the "DNA of a banker", analytics can help identify value in a banking portfolio. He will also discuss the challenges of implementing such an approach and the increasing importance of predictive analytics in previously untouched areas of banking.

Speaker: Gaurav Gaur, Director, Product Management, The Royal Bank of Scotland

[ Top of this page ] [ Agenda overview ]


Track 1: Nonprofit Fundraising
Case Study: JustGiving
Analytics at JustGiving - Using data to unlock and increase individual generosity

Many organisations encounter the same problem of generating large volumes of valuable data but lacking the ability or time to use it to drive business objectives. JustGiving was no exception, but we are now using data and analytics to enable people and communities to raise extraordinary amounts of money for the causes they care about. This presentation explains how JustGiving is using a strategy based on predictive analytics that is set to take JustGiving's total fundraising for charities from £700 million to one billion pounds and beyond. JustGiving will use predictive models in conjunction with other statistical techniques to analyse our large data repository of past user behaviour. This analysis will allow us to build behaviour profiles of new and existing customers which we can use to predict the likely transactional behaviour, charitable preferences, propensity to give and much more.

Speaker: Mike Bugembe, Head of Analytics, JustGiving

[ Top of this page ] [ Agenda overview ]

Track 2: Operational Analytics
Presentation from : HM Revenue & Customs
From Analysis to Action: Enabling strategic change through operational analytics

Operational analytics links front-line decision-making to strategic planning and performance measurement. Its effective design can turn operational IT systems from costly barriers to effective enablers of business change. Chris will talk about his experiences in deploying operational analytics in a debt collection environment and developing a professional approach to analytics service delivery. Tom will conclude on a direction for professionalism in Predictive Analytics.

Speaker: Dr Christopher Hemingway, Head of Analytics, HM Revenue & Customs
Speaker: Tom Khabaza, Chairman, Institute of Data Miners

[ Top of this page ] [ Agenda overview ]


Session Break


Track 1: Demand Forecasting
Case Study: OTTO
Inventory improvement with Predictive Analytics at the world's 2nd largest online retailer OTTO

Retail organisations face extreme challenges in keeping their stocks as small as possible and at the same time never running out of stock for their customers. Perfect stock planning reflects real competitive advantage and increases profitability. Otto Group, a multi-channel retailer and the world’s 2nd largest online retailer after Amazon, improved its short-, mid- and long-term sales forecasts through predictive analytics software combined with a neuronal network. This presentation shows you how the Otto Group achieved prediction improvements of between 20-60%. It provides insights into how the Otto group raises its customer satisfaction by optimising their inventory throughout the different seasons. The advanced analytics findings enabled Otto to reduce unsold goods at the end of the season to a perfect minimum. Further areas of improvements through predictive analytics are the optimisation of multi-channel ordering forecasts with predictive analytics and an online recommendation engine for targeting.

Speaker: Michael Sinn, VP Buying Division, OTTO

[ Top of this page ] [ Agenda overview ]

Track 2: Customer Insights
Case Study: The Economist & Netto Bank
New Directions in Customer Intelligence

This session summarises a series of Customer Intelligence projects demonstrating fast and robust techniques to discover new insight about customers’ value, behaviour and needs. This goes far beyond simple Recency, Frequency, Total value and demographic segmentation. Examples are purposely drawn from industries where there are not a lot of customer transactional records, such as subscription publishing at The Economist and private banking at Netto Bank, to show how customer intelligence is being created within low customer-transaction BtoC AND BtoB industries.

Speaker: Rosaria Silipo, Data Mining Consultant

Speaker: Phil Winters, Senior Managing Partner, CIAgenda

[ Top of this page ] [ Agenda overview ]


Track 1: Demand Forecasting
Case Study: ENERGEX
Spatially Simulating Electricity Demand Growth

With the help of external resources, ENERGEX developed a spatial simulation which forecasts 20 years of electricity demand growth in South East Queensland, Australia. Techniques such as agent-based urban simulation, machine learning and customer segmentation were employed to forecast over 250 variables to a 400x400m grid level. The spatial simulation developed will likely form part of ENERGEX's next regulatory submission which determines ENERGEX's funding. In this session, Jared will talk more about the modelling approach and how ENERGEX is using the simulation.

Speaker: Jared McKee, Customer Integration Specialist, Energex

[ Top of this page ] [ Agenda overview ]

Track 2: Social Data; Financial Services
Case Study:
How Do Peers Evaluate Peer-to-Peer Lending?

Is peer-to-peer lending different from the traditional bank-initiated lending? How do the customers/lenders think and act? We used a comprehensive p2p data and found some interesting behavioural and analytical results on social interactions.

Speaker: Aaron Lai, Senior Manager of Marketing Analytics, Blue Shield of California

[ Top of this page ] [ Agenda overview ]


Refreshment Break

Case Study: GlaxoSmithKline
Predictive Analytic Patient Recruitment and Drug Supply Modelling in Clinical Trials

Patient recruitment and drug supply stage is a well-recognised bottleneck for the design and monitoring of clinical trials. A large proportion of trials fail to recruit in time and a huge proportion of drugs is wasted. New analytic methodologies for predictive patient recruitment modelling over time (with mean and confidence bounds) and for risk-based drug supply modelling has been developed. Software tools (using R and RExcel) have been created. Both tools are now implemented for all large Phase III trials and already enabled significant cost savings and benefits in R&D GlaxoSmithKline. We are not aware of similar tools used by other pharma companies.

Speaker: Vladimir Anisimov, Senior Director, GlaxoSmithKline

[ Top of this page ] [ Agenda overview ]


Refreshment Break

[ Top of this page ] [ Agenda overview ]

Social Data; Text Analytics
Case Study: BBC
Data Mining for Social Moderation

The BBC (the British Broadcasting Company), a known media force in Europe, implemented an in-database data mining solution for their public-facing website. The BBC encourages users to post comments and become part of their social media community. In this project, the BBC needed a way to improve the social moderation of millions of posts on thousands of forums. This presentation covers integration within a leading database platform, and final model results evaluation. We also discuss cost efficiencies. More organisations will use enterprise solutions to handle high-volume social networking.

Speaker: Mark Tabladillo, Mentor, SolidQ
Speaker: Francisco Gonzalez, Mentor, SolidQ

[ Top of this page ] [ Agenda overview ]


Reception followed by Data Driven Business Week Awards Ceremony

[ Top of this page ] [ Agenda overview ]

Day 2: Thursday, 1 December, 2011


Registration & Coffee

[ Top of this page ] [ Agenda overview ]

Multiple Case Studies: U.S. DHS, SSA. Text Mining: Lessons Learned

Text Mining is the "Wild West" of data mining and predictive analytics - the potential for gain is huge, the capability claims are often tall tales, and the "land rush" for leadership is very much a race. In solving unstructured (text) analysis challenges, we found that principles from inductive modelling - learning relationships from labeled cases - has great power to enhance text mining. Dr. Elder will highlight key technical breakthroughs discovered while working on projects for leading government agencies, including :

  • Prioritising searches for the Dept. of Homeland Security
  • Quick decisions for Social Security Admin. disability
  • Disease discovery for the Dept. of Homeland Security

Speaker: John Elder, Ph.D., Elder Research, Inc.

[ Top of this page ] [ Agenda overview ]


Refreshment Break

Expert Panel: Wise Enterprise: Best Practices for Managing Predictive Analytics

Your company is trigger-happy for predictive analytics, and there’s plenty of excitement, momentum and public case studies fueling the flames. Are you destined for success or disappointment? Is it a sure-fire win to gain buy-in for a promising analytics initiative, equip your most talented practitioners with a leading solution, and pull the trigger?

This panel of leading experts will address the holistic view. What are the most poignant and telling failures in the repertoire, and where is the remedy? Beyond the management of individual analytics projects, what enterprise-wide communication processes and other best practices provide best security against project pitfalls? Stay tuned for big answers to these big questions.

Speaker: Uwe Weiß, CEO, Blue Yonder

Speaker: Colin Linsky, Principal Analytical Consultant, IBM

Speaker: Tom Khabaza, Chairman, Institute of Data Miners

[ Top of this page ] [ Agenda overview ]


Lunch / Exhibits
Room: Expo Hall

Special Featured Session: Uplift Modelling
Uplift Modelling: You Should Not Only Measure But Model
Incremental Response

Most marketing analysts understand that measuring the impact of a marketing campaign requires a valid control group so that uplift (incremental response) can be reported. However, it is much less widely understood that the targeting models used almost everywhere do not attempt to optimise that incremental measure. That requires an uplift model.

This session will build on Eric Seigel’s Keynote, diving deeper into why a switch to uplift modelling is needed, illustrating what can and does go wrong when they are not used and showing the hugely positive impact they can have when used effectively. It will go on to explain how uplift models can actually be built and assessed.

Speaker: Neil Skilling ,Director Product & Solutions, Pitney Bowes Business Insight

[ Top of this page ] [ Agenda overview ]


Crowdsourcing Predictive Analytics
Predicting Retail Customer Behaviour

We all know that everyone has to go grocery shopping at some point but predicting exactly when a customer will next visit their local grocery, and how much they will spend, is a thorny challenge. Armed with a year's data on over 100,000 visits, dunnhumby hosted a competition to increase predictive accuracy on future visits with impressive results. This session will introduce the idea of predictive modeling competitions using the dunnhumby Shopper Challenge as a case study.

Speaker: Giles Pavey, Head of Retail Solutions, dunnhumby
Speaker: Mehul Patel, Chief Operating Officer, Kaggle

[ Top of this page ] [ Agenda overview ]



Project Management for Analytics
A Proposed Methodology for Integration of Predictive Analytics in Organisations

Stakeholders required a standard process around predictive analytics (PA) initiatives in order to ascertain what could be done with PA and how PA relates and how it is relevant to their business. We present a new, tested standard, InterActive8, which reflects that PA integration goes qualitatively beyond merely successfully completing the hands-on portion of a PA project. This standardised process builds on the "Plan, Do, Study, Act" (PDSA) cycle proposed by statisticians Shewhart and Deming, which had tremendous impact on the "continuous improvement" movement of recent decades. InterActive8 focuses on problem-solving action (business relevance), interdisciplinary dialogue between analysis and action, and managing organisational complexities around value longevity that arise, for example, upon changes in business sponsorship, and inertia against entailed operational changes.

Speaker: Uros Bole, Principal, Pikarp d.o.o.

Speaker: Jure Zabkar, PhD, Al Lab, University of Ljubljana

[ Top of this page ] [ Agenda overview ]

Discovery of Technology Trends
The Science of Science: Technical Opportunity Discovery Model-driven Analytics System

This presentation shows how we developed an analytics system which is able to determine current and future stages in a given technology lifecycle using our technical opportunity discovery (TOD) model. The TOD model-driven system named InSciTe Advanced aims to support systemic decision-making processes, especially focused on R&D strategy planning. It combines text mining and Semantic Web technologies to facilitate extraction of technical terms and their relations and furthermore interoperable integration of heterogeneous literatures. InSciTe Advanced provides multi-target solution functions to find emerging technologies in all science and technology areas and single target solution to show multi-faceted related information for each technology. After developing internal service for tens of millions of scientific journal papers and patents, it has recently been applied to the Defense Agency for Technology and Quality (DTaQ) of Korea to analyse millions of defence documents and to generate technology summary reports automatically.

Speaker: Hanmin Jung, Chief Researcher, KISTI

Speaker: Seungwoo Lee, Ph.D., Senior Researcher, KISTI

[ Top of this page ] [ Agenda overview ]


Refreshment Break


Case Study: Hewlett-Packard
Churn/Threat Prediction Analysis in B2B by Value Movement Analysis

Identify how valuable an account is to HP and what is the propensity of that account to purchase HP's offerings. To overcome the challenges of data paucity and lack of application of standard industry analyses – we assign an RFM score to each account across multiple, successive time periods. This predicts the score of the account and subsequent categorisation into – Valuable, Churn, Growth – and can therefore be used to prioritise account focus and resultant sales-force action. The score and account categorisation is validated across time by referring to live data and the model is refreshed/refined based on past prediction errors

Speaker: Subhamitra Chatterjee, Marketing Analyst, Hewlett-Packard

Speaker: Dwipashray Niyogi, Analytics Consultant, Hewlett-Packard

[ Top of this page ] [ Agenda overview ]

Supply Chain Management
Case Study: United Group Holdings
Value Proposition Segmentation (VPS) Method

In today’s competitive market, effective management of customer relations lies in the ability to optimise the dual creation of firm (shareholder) and customer value. Accordingly, the challenge for many companies is to be able to understand and differentiate heterogeneous customers by their needs to deliver the wining value proposition profitably. Our proposed VPS model addresses the basic managerial concern of balancing relationships from both the seller’s (customer loyalty) and the buyer’s (customer benefit), by considering both the service provider’s financial performance (i.e. customer value to the firm) and the value customers receive from the provider’s offerings (i.e. customer benefit).

Speaker: Amjad Zaim, CEO, Cognitro Analytics

[ Top of this page ] [ Agenda overview ]

Friday, December 2, 2011

Full-day Workshop
The Best and the Worst of Predictive Analytics: Predictive Modelling Methods and Common Data Mining Mistakes

  • Workshop starts at 9:00am
  • First AM Break from 10:00 - 10:15
  • Second AM Break from 11:15 - 11:30
  • Lunch from 12:30 - 1:15pm
  • First PM Break: 2:00 - 2:15
  • Second PM Break: 3:15 - 3:30
  • Workshops ends at 4:30

Instructor: John Elder, Ph.D., CEO & Founder, Elder Research, Inc.

Share |

PAW Highlights Video

Exhibit Hours


Blue Yonder Pitney Bowes

AnalyticbridgeAnalyticBridge Customer Think

improvement & innovation Institute of Data Miners KDnuggets Top Seos Visibility Magazine TECHWORLD

Check out the official PAW Blog Partners
© 2019 Predictive Analytics World | Privacy
Produced by Prediction Impact, Inc. and Rising Media, Inc.

Predictive Analytics Company           Predictive Analytics Event Producer