Transcript StephanSorger.com

Product and Service Analytics
Disclaimer:
• All logos, photos, etc. used in this presentation are the property of their respective
copyright owners and are used here for educational purposes only
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.1
Conjoint Analysis
Conjoint Analysis for Tablet Device
Attributes
- Operating system, screen size, battery life
Attribute Levels
- Screen Size: 5 inch, 7 inch, 10 inch
Bundles
- Different combinations of attributes
Conjoint Analysis
- Technique to examine trade-offs consumers
make to understand their preferences
Part-Worths
- Values placed on particular attributes
Profiles
- Specific bundles preferred by segments
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.2
Conjoint Analysis: Process
Prepare
for
Conjoint
Collect
Preference
Data
Code Data
for
Analysis
Calculate
Attribute
Part-Worths
Apply
Conjoint
Results
Step
Description
Prepare for Conjoint
Identify evaluation attributes
Select levels for each attribute
Form bundles (candidate “products”)
Get Preference Data
Survey consumers for their preferences
Code Data
Prepare data for analysis by coding it
Calculate Part-Worths
Calculate preference for each attribute
Apply Results
Interpret to assess market size and segmentation
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.3
Conjoint Analysis: Process
Conjoint Analysis Preparation
Identify
Evaluation
Attributes
Select
Attribute
Levels
Form
Candidate
Bundles
Topic
Description
Identify Evaluation Attributes
Review available consumer evaluation sources
General sources: Amazon.com, Epinions.com, etc.
Specialty sources: CoffeeGeek, Home-Barista
Conduct survey of top attributes (next slide)
Select Attribute Levels
Apply knowledge gained from study of category
Form Candidate Bundles
Combine various attribute levels to form bundles
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.4
Conjoint Analysis: Process
Example: Acme Espresso Machines
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.5
Conjoint Analysis: Process
Acme Espresso Machine Attribute Levels
Attribute Levels for Non-Numeric Values
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.6
Conjoint Analysis: Process
Candidate Bundles, also known as “Cards”
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.7
Conjoint Analysis: Process
Data Collection Techniques
Pairwise
Comparison
Rank
Ordering
Rating
Scale
Topic
Description
Pairwise Comparison
Respondents compare pairs of options
Advantage: Respondents find easy to evaluate
Disadvantage: Requires many comparisons
Rank Ordering
Respondents place options in rank order: 1 – 100
Advantages: Fast
Disadvantages: Respondents find it difficult
Rating Scale
Respondents rate each option independently
Advantages: Works well with Excel
Disadvantages: Must provide rating scale
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.8
Conjoint Analysis: Process
Pairwise Comparison
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.9
Conjoint Analysis: Process
Pairwise Comparison
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.10
Conjoint Analysis: Process
Rating Scale
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.11
Conjoint Analysis: Process
Sample Respondent Preference Results
Sample Respondent Segmentation Identification Results
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.12
Conjoint Analysis: Process
Coding Process
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.13
Conjoint Analysis: Process
Sample Respondent Results, Coded into Binary for Easier Machine Computation
Binary Coding with Three Levels
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.14
Conjoint Analysis: Process
Sample Respondent Results, with Redundancies Removed
Remove redundancies to prevent linear dependency problems
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.15
Conjoint Analysis: Process
Excel
Regression
…
Home
…
Data
Input Y Range
Data Analysis
A
B
C
D
E
F
G
OK
Input X Range
x Labels
Constant is Zero
x Confidence Level:
Launching Data Analysis in Excel
95
%
Entering Data in Regression Dialog Box
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.16
Conjoint Analysis: Process
Microsoft Excel Regression Results
Preference = Constant + A1 * Speed 1 + A2 * Capacity 1 + A3 * Price 1
Preference = 2.0 + 1.75 * Speed 1 - 0.75 * Capacity 1 + 1.25 * Price 1
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.17
Conjoint Analysis: Process
Conjoint Application
Market
Segmentation
Market
Simulation
Topic
Description
Market Segmentation
Correlate conjoint data with segmentation data
(Demographic, Geographic, Behavioral, Psychogr.)
High part worth utility for speed  “Used at work”
Market Simulation
Collective voice of hundreds of potential customers
Simulate market reception to new machine
First choice rule: Respondents choose 1 product
Market share: % of respondents with high utility
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.18
Decision Tree Models
Establish
Decision
Choices
Gather
Relevant
Data
Calculate
Random
Node Values
Calculate
Decision
Node Values
Select
Winning
Alternative
Topic
Description
Decision Choices
List out alternatives
Relevant Data
Gather data for each alternative
Random Node Values
Calculate values at random nodes
Decision Node Values
Calculate values at decision nodes
Uses results from random node calculations
Winning Alternative
Select alternative with highest net expected value
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.19
Decision Tree Models
Development Project Decision
Develop New Product
A. Use Standard Budget
Strong
Average
Poor
Decision Node
Random Node
Scenario
B. Use Reduced Budget
Strong
Average
Poor
C. Enhance Existing Product
Strong
Average
Poor
Typical Development Project Selection Scenario
Step 1: Establish Decision Choices
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.20
Decision Tree Models
Step 2: Gather Relevant Data: Choice 1, Standard Budget
Step 2: Gather Relevant Data: Choice 2, Reduced Budget
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.21
Decision Tree Models
Step 2: Gather Relevant Data: Choice 3, Develop Existing Product
Step 2: Gather Relevant Data: Costs for Each Alternative
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.22
Decision Tree Models
Step 3: Random Node Value: Alternative 1: New Product, Standard Budget
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.23
Decision Tree Models
Step 3: Random Node Value: Alternative 2: New Product, Reduced Budget
Step 3: Random Node Value: Alternative 3: Enhance Existing Product
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.24
Decision Tree Models
Development Project Decision
Net EV = $126,000
EV = $326,000
Cost = $200,000
Net EV = $126,000
Develop New Product
A. Use Standard Budget
Strong
Average
Poor
EV = $152,000
Cost = $100,000
Net EV = $52,000
B. Use Reduced Budget
Strong
Average
Poor
EV = $63,000
Cost = $40,000
Net EV = $23,000
C. Enhance Existing Product
Strong
Average
Poor
Step 4: Calculating Decision Node Values
Step 5: Select Alternative with Highest Net Expected Value
Winner: “Develop New Product, Standard budget”
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.25
BCG Matrix: Product Portfolio Allocation
List
Products
Enter
Data
Assign
Rating
Assign
Status
Allocate
Resources
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.26
Product/Service Metrics
Product/ Service Sales Input Table: Total Revenue by Month, Products A, B, and C
Product/ Service Sales Input Table: Revenue in Different Markets by Month, Product A
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.27
Product/Service Metrics
Product A: Steady Rise
Product B: Declining
Product C: Seasonal
January
Time
December
Product/ Service Sales: Revenue Trends
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.28
Product/Service Metrics
Revenue
A
A
B
A
B
B
C
Market 1
C
C
Market 2
Market 3
Product/ Service Sales: Market Adoption
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.29
Product/Service Success Quadrants
Product A
Gross
Margin
Niche Stars
Super Stars
Product B
50%
Concern Areas
Mass Market
Product D
Product C
Revenue
Product/ Service Profitability: Product Success Quadrant Tool: Graphical Format
Adapted from product profitability analysis tool by Demand Metric; Used with permission
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.30
Product/Service Success Quadrants
Illustrative method to group products/ services by profitability
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.31
SEM Attribute Preference Test
Google
Left
Nav.
A Vacuum Carpets Fast
Turbo-Vortex design
Delivers 2x the suction!
www.acmevacuum.com
Search Box
Featured Ads
Ads
Organic
Search
Results
Pay Per Click Ads
such as Google AdWords
B
Hey Allergy Sufferers!
Hyper-HEPA filter
Removes 1-micron particles
www.acmevacuum.com
C
Vacuum Drapes Easily
EZ-DRAPE attachment
Cleans curtains with ease!
www.acmevacuum.com
Test
A
B
C
Clicks
240
300
4
Buys
12
2
0
Apply SEM to test attribute preferences
© Stephan Sorger 2013. www.StephanSorger.com; Marketing Analytics: Product Analytics 7.32