Models Collecting Dust? How to Transform Your Results from Interesting to Impactful

Models Collecting Dust?

James TaylorCEO

©2012 Decision Management Solutions 2

Your presenter – James Taylor

CEO of Decision Management Solutions

Works with clients to improve their business by applying analytic technology to automate & improve decisions

Spent the last 9 years championing Decision Management and developingDecision Management Systems

ChallengesOr why your models are

gathering dust


Knowing is not enough

Those who know first, win

Those who ACT first, winProvided they act intelligently

It’s hard to communicate your value

5©2012 Decision Management Solutions

Time to deploy models matters


Cottage industries don’t scale


Challenges with predictive analytics

Actions are needed not just predictionsBegin with the decision in mind

The business does not understand analyticsA business context for analytics

Analytic models age quicklyCottage industries don’t scale

Industrialize your analytic processes


Begin with the decision in mind

Decision Management is the proven approach used to

manage decisions and apply predictive analytics effectively



3 Steps to Decision Management

DiscoverBuildImprove


What is a decision?

Data is gathered, consideredA choice or selection is madeThat results in a commitment to action


Different kinds of decisions

Strategic Decisions• Few in number, large impact• Should we acquire this company or exit this market?

Tactical Decisions• Management and control, moderate impact• Should we re-organize this supply chain, change risk management approach?

Operational Decisions• Day-to-day decisions that affect one transaction or customer• Best offer for this customer ?How risky is this loan? Is this claim fraudulent?


Decisions are the focal point for risk

Risk is not acquired in “big lumps” but one transaction at a time


Decisions maximize customer value

Risk Opportunity


Three kinds of analytic decisions

Fraud


Analytics power operational decisions

How do I…prevent this customer from churning?convert this visitor?acquire this prospect?make this offer compelling to this person?identify this claim as fraudulent?correctly estimate the risk of this loan?

It’s not about “aha” momentsIt’s about making better operational decisions

Case: RetailerBusiness challenges Solution Benefits

Grocery chains are battling for market shareCustomer loyalty is essential for growthLoyalty to the brand, not a single store format

Tailored promotions integrated with loyalty programIntegrated system from back office to point of saleConsistently compelling offers across channels

Increased revenueDeep knowledge of customers across formatsMore effective promotional campaigns


A business context for decision making


Candidate decisions

Determine if a customer is eligible for a loan

Calculate the discount for an order

Assess the risk of a transaction

Select the terms for a deal

Choose which claims to Fast Track

These are decision wordsThe system must answer a question each time


Target decision-making on KPIs

The link between decisions and KPIs is critical

Operational decisions affect KPIs

KPIs measure operations

Strategy defines KPIs


Decision to KPI mapping

KPIs impacted by improvement in decision

CustomerChurn

Customer Service Calls

Losses Retention Budget

KPI 5

What retention offer should be made? ↑ ↑ ↓ ?What initial price should be offered? ↑ ↑Should an intervention call be made? ↓ ↓Decision 3 ?…


Decompose the decisioning

What is required to make decision?Guidelines, policy documentsHuman expertiseRegulationsExisting system logic

Data describing the caseExternal reference data

Predictive Analytic ModelsData Mining Results

The results of other decisions


Decisions provide context

ProcessesActivities require decisions

EventsTrigger decisions

SystemsImplement decisions

Organizational UnitsMake decisionsOwn decisionsAre impacted by decisions

Know which business processes will be improved by your analytics

Know when your analytics might be calculated

Know how you will have to deploy your analytics

Know who cares about your analytics and who will have to believe them

Case study: InsurerBusiness challenges Solution Benefits

Use analytics to improve underwriting

Embed analytics in claim processing application

Model and decompose decisionsMap decisions to systems and organizations

Find the decisions that could be impacted and refocused analytic effortConstrained analytic effort to ensure successful implementation

Industrialize Analytics


Embed Predictive Analytics

Decision

?

?

In-database Modeling

In-database Scoring

Model Tuning

Data ManagementData PreparationData Visualization & AnalysisModelingModel ValidationDeployment and ScoringModel MonitoringModel TuningRepository


Analytic Insight Management


Build decision-making components

Operational Systems

AnalyticSystems

Decision

Case study: Cable TVBusiness challenges Solution Benefits

1.2M householdsMany single-product householdsWhole industry suffers from low loyalty and 20%+ customer churnIncreasing competition and changing regulations

Predictive analytics to predict churn, cross-sellBusiness rules use analytics and data to drive dynamic scriptsEmbedded in call center application to improve decision making

13-18% cross-sell hit rate on average

Up to 40% cross-sell success rate for some

Teams using the scripts have more sales

Reduced churn by 20-30%


Some Pitfalls and how to avoid them.


Operational decisions are at the center

Business

Decision


Don’t just create a decision point


Create Continuous improvement

Case: State dept of taxationBusiness challenges Solution Benefits

Paper tax returns increased costs and slowed responsesInformation system silosManual fraud detection and return review

Single central taxpayer databaseIntegrated systemSophisticated real-time predictive analytics

Recovered millions of dollars from dubious tax returnsIncreased collection of unpaid taxesDecreased number of questionable returnsIncreased customer satisfaction


Questions?


Decision Management SystemsWhat if you could make your systems active participants in optimizing your business?

What if your systems could act intelligently on their own?

Decision Management Systems can do all that and more. This book shows how to integrate operational and analytic technologies to create more agile, analytic, and adaptive systems.

Discount Code: TAYLOR4389

For more information about this new release, visitwww.decisionmanagementsolutions.com/book

In a Predictive Enterprise, analytics are…Pe

rvas

ive

Pred

ictiv

eA

ctio

nabl

e

Used in every transactionAt the point of contact/deliveryIn operational decision making

From reporting to prediction and forecastingData miningPredictive analytics and scoring

Decisions being made, actions being takenDecision Management SystemsDecision Support Systems



Begin with the Decision in mind

Find the decisions that matter to your business and model them

DiscoverBuildImprove

Create an Analytic Factory


Thank You

43

James Taylor, [email protected]

Models Collecting Dust? How to Transform Your Results from Interesting to Impactful

Business

decision management

tactical decisions management

decision wordsthe system

risk management approach

decisions processes

decisions andkpis

aboutmake decisions

candidate decisions