Page 1
A Conceptual Framework for Managing Customer
Experience and AnalyticsTMF TAW
Lisbon
January 2010
Dr. Lorien Pratt
Quantellia, LLC
[email protected]
Blog: www.lorienpratt.com
Page 2
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Page 3
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
MCEAnalyti
cs
Agenda
Page 4
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
MCE
Agenda
Page 5
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Holistic MCE Models
Page 6
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Holistic MCE Models
Page 7
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Holistic MCE Models
Page 8
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Holistic MCE Models
Page 9
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Holistic MCE Models
Page 10
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Key Factor Analysis (TR148, TR149)
Process
$
People
Touchpoint
Loyalty Profitability
Page 11
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
MCEAnalyti
cs
Agenda
Page 12
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Analytics
Agenda
Page 13
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Examples of Analytics in MCE
• Measuring order fallout
• Predictive resource allocation
• Installation process optimization
- Offer design and analysis
- Customer segmentation / marketing
- Upsell triggering
- Serviceability analysis
- Lifetime Customer Value prediction
- Recommendations
- Personali-zation
- Advertising- Direct
Marketing- Retail
Placement
• CDR analysis
• Payment, credit, cash flow forecasting
• Leakage identification
• Personalization
• Fraud identification
• Parental Control
• VOD Purchasing Behavior
• Clickstream analysis’
• Relationship building / loyalty
Acquisition
Fulfillment Usage Support Optimi-
zation
• SLA Analysis
• Multi-level support
• Retention• Next-best
offer
Page 14
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Decision Engineering
Adaptive Analytics
Predictive Analytics
Reporting
Data Management (including collection, ETL, deduplication, aggregation, correlation, data migration, data quality, data
modeling)
Page 15
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Decision Engineering
Adaptive Analytics
Predictive Analytics
Reporting
Data Management (including data migration, data quality,
data modeling)
Page 16
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
One-slide predictive/adaptive analytics overview
Will this customer churn?
Yes/No data: If customer has an open trouble ticket: Yes, otherwise:
No
Real-Valued: If customer age < 30: Yes, otherwise: NoCombination: If customer age <30
AND has an open trouble ticket: Yes, otherwise: No
Linear Combination: If 2.3 x Age + 4.4 x Income > 40: Yes,
otherwise: No
Predictive Analytics: Obtain these numbers by analyzing historical
data
Adaptive Analytics: Update your historical data, and re-derive the numbers periodically to take changing situations into account.
Nonlinear Analytics:
age
Income vs.
age
Income
Pattern
4.1 2.1 3
Page 17
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Decision Engineering
Adaptive Analytics
Predictive Analytics
Reporting
Data Management
(including data migration,
data quality,
data modeling)
Page 18
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Decision Engineering: Unifies manual and automated decision making
• Measuring order fallout
• Predictive resource allocation
• Installation process optimization
- Offer design and analysis
- Customer segmentation / marketing
- Upsell triggering
- Serviceability analysis
- Lifetime Customer Value prediction
- Recommendations
- Personali-zation
- Advertising- Direct
Marketing- Retail
Placement
• CDR analysis
• Payment, credit, cash flow forecasting
• Leakage identification
• Personalization
• Fraud identification
• Parental Control
• VOD Purchasing Behavior
• Clickstream analysis’
• Relationship building / loyalty
Acquisition Fulfillment Usage Support Optimi-zation
• SLA Analysis
• Multi-level support
• Retention• Next-best
offer
Many decisions are made manually. Why:- When the future is not
like the past, analytics is not enough
- Missing data- Uncertain data- Changing situation- Complex situation
We cannot wait for complete data to support MCE decision making.
Where should I invest my MCE dollars?
Page 19
Data
Data
Gap
www.quantellia.com
The Decision Support Problem
Page 20
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Systematic Decision Making Problems*
• “We focus on only one measure, when there are really multiple objectives.”
• “We make decisions that assume a predictable unchanging future.”
• “Our focus is on short-term goals, ignoring long-term ones.”
• “We are unable to reason about long cause-and-effect chains.”
• “We ignore intangibles like morale, reputation, trust, and brand.
• “We plan for only a single future scenario when radically different courses of action may be appropriate, depending on how the future unfolds.” Revenu
e Community Service
Cost
“Five years from now, the market for our product will have grown by 30%”
“I can barely plan for next quarter, how can I think about the future, too?”
Reduce Time We Spend on Customer Care Telephone Calls
Lower Customer Care Costs
Improved Contribution Margin
Unhappier Customers
Reduced Knowledge of our Customers
Greater Customer ChurnSmaller Profits
Brand
*High Performance Decision Making. Pratt and Zangari, 2009
Page 21
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Tactical CEM Decision Engineering Process
Define Objectives
and Specifications
Analyze Data Needs &
Availability
Design Decision Model
Determine Inputs and Outcomes
Clearly specify:• Terminology.• What is to
be achieved.• What are
the constraints.
Is historical data relevant? Or will this initiative change internal or external behavior to make past data misleading?
Construct the appropriate decision model:• Extrapolate
from past data, or
• Model new system
Use decision model to determine execution parameters and baseline performance metrics.
Strategic Decision Engineering
Define Objectives and Specifications
Analyze Data Needs & Availability
Design Decision Model
Determine Inputs and Outcomes
Page 22
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Strategic CEM Decision Engineering Process
Define Objectives and Specifications
Analyze Data Needs & Availability
Design Decision Model
Determine Inputs and Outcomes
• Compare outputs of decision design process for different alternative courses of action.
• Determine which option best meets the company’s business needs.
• Begin Execution.
Execution
Page 23
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
CEM Execution
Define Objectives and Specifications
Analyze Data Needs & Availability
Design Decision Model
Determine Inputs and Outcomes
Measure effectson customer behavior, costs
and revenues.
Customers reactto external effects ofnew initiative.
Create/updateinitiatives based on analysis of models and data
New initiatives have internal and external effects
Unexpected effects may require re-plan.
Page 24
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
Design versus monitoring
KPI #1
• Like automobile design• Key competency: being
able to understand how the system will work
• Key competency: using judgment where data is missing
• Like monitoring a working vehicle
• Key competency: detecting problems accurately and quickly
• Key competency: diagnosis
Page 25
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
√
√√√
√√
A comprehensive analytics strategy
addresses many MCE issues
Page 26
U.S. Patent Pending. All rights reserved. Copyright © 2010 Quantellia Inc.
MCEAnalyti
cs
Conclusion
• MCE has moved from individual touchpoints to a holistic approach
• Data management and analytics support several parts of this process• To be “Actionable”, CEM data must support decisions
• These decisions are tactical and strategic, and can include investment / ROI decisions
• Data must support both manual and automated decision making
• When the future is not like the past, a “computer aided decision design” approach is helpful
Page 27
THANK YOU
Dr. Lorien Pratt
[email protected]
+1 650 943 2444
Blog: www.lorienpratt.com