understanding the data behind pricing - dan barlow, epacube
DESCRIPTION
Distribution recognizes the need to develop talent for pricing within their organizations. Owning this core competency is now critical for survival. Many distributors have historically outsourced this capability, but are now ready to adapt this skill set internally and own the most important component to their customer relationship. In this session, we will discuss the transactional datasets used for pricing analysis and provide an understanding to the importance of blending knowledge with data to derive accurate pricing for customers. We will point out potential pitfalls in pricing that lead to commercial issues, and discuss the use of segmentation to derive contextual performance models.TRANSCRIPT
Understanding the Data Behind Pricing
Dan Barlow, epaCUBE
Agenda Understanding the Data Behind Pricing
Composition of a Price Matrix
The Science Behind Pricing
Dealing With Bad Behavior
Blending Art With Science
Bad Segments Corrupt Good Math
PRODUCT SEGMENTATION • Proper Grouping of Like Material • Prevents Inconsistencies in Pricing • Typically Provided By The Vendor
Begins With Segmentation Composition of a Price Matrix
CUSTOMER SEGMENTATION • Proper Grouping of Like Customers • Drives Consistency in Strategy • Typically Provided By The Sales Team
Customer Segment
A Sales Rep or Management Decides What Other Customers This Particular Customer Should Be Priced Like
Product Segment
Product Segments Are Usually Provided By The Vendor Or Grouped By A Commodity Status
Strategic Outcome
The Matrix Price Is Derived Off Some Sales History, but More in Line With What Sales Feels Is Reasonable
Strategic Matrix Design Typical Nature Behind Most Matrix Creations
The Price Matrix A Relationship Of Pricing By Customer Groups to Product Groups
o Based Heavily on Strategy, and Simple Math
o Analysis Can Be Difficult
o Customer Segments Usually Lack Structure
o Limited Visibility To See Performance Concerns
o Do Not Know Who To Change and By How Much
o Lack Agility to Hide In Vendor Updates
o They Simply Age With Limited Upkeep
The Science Behind Pricing The Math That Backs Up The Business Strategy
o Statistical Evaluation Of Typical Sell Price Performance
o Used To Derive:
Typical Performance
Suggested Price Points
How Things Compare
What is Too High and What is Too Low
Defining The Middle
MEAN MEDIAN MODE
13, 18, 13, 14, 13, 16, 14, 21, 13
= (13 + 18 + 13 + 14 + 13 + 16 + 14 + 21 + 13) / Count = (135) / 9
= 15
13, 18, 13, 14, 13, 16, 14, 21, 13 13, 13, 13, 13, 14, 14, 16, 18, 21
= 14
13, 18, 13, 14, 13, 16, 14, 21, 13 (13, 13, 13, 13)(14, 14) 16, 18, 21
= 13
Definitions
MEAN MEDIAN MODE
= 15
= 14
= 13
Represents the Average of the Data.
Represents the Centermost point in the Data.
Represents the Most Repeated of the Data.
13, 18, 13, 14, 13, 16, 14, 21, 13
Standard Deviation
A Numerical Representation of How Well The Mean Represents the Data Points
So…How Far Is Each Data Point
In Comparison to The Mean?
13, 18, 13, 14, 13, 16, 14, 21, 13
MEAN
= (13 + 18 + 13 + 14 + 13 + 16 + 14 + 21 + 13) / Count = (135) / 9
= 15
Going The Distance
13 - 15 = -2 18 - 15 = 3 13 - 15 = -2 14 - 15 = -1 13 - 15 = -2 16 - 15 = 1 14 - 15 = -1 21 - 15 = 6 13 - 15 = -2
If You Add These Together, You Get 0, So You Square The Values!
MEAN
= (13 + 18 + 13 + 14 + 13 + 16 + 14 + 21 + 13) / Count = (135) / 9
= 15
13, 18, 13, 14, 13, 16, 14, 21, 13
Going The Distance MEAN
= (13 + 18 + 13 + 14 + 13 + 16 + 14 + 21 + 13) / Count = (135) / 9
= 15 Total = 64, Then…
You Divide 64 / (Count – 1)
So, 64/9-1 or 64/8
= 8
Lastly, We Take The Square Root of 8
= 2.83
13 - 15 = -22 = 4 18 - 15 = 32 = 9 13 - 15 = -22 = 4 14 - 15 = -12 = 1 13 - 15 = -22 = 4 16 - 15 = 12 = 1 14 - 15 = -12 = 1 21 - 15 = 62 = 36 13 - 15 = -22 = 4
13, 18, 13, 14, 13, 16, 14, 21, 13
How Far Is Each From Mean?
-2.83 -2.83 +2.83 +2.83 +2.83 -2.83
13, 18, 13, 14, 13, 16, 14, 21, 13 MEAN
15
How It Lines Up
-2.83 -2.83 +2.83 +2.83 +2.83 -2.83
13, 18, 13, 14, 13, 16, 14, 21, 13
9.34 12.17 17.83 20.66 23.49 6.51
MEAN
15
Empirical Rule
68%
95%
99%
13, 18, 13, 14, 13, 16, 14, 21, 13 MEAN
PRODUCT SEGMENTATION • Dissimilar List Price Points
• Different Value Propositions • Wide Margin Spreads
Bad Data In…Well, You Know… Bad Segments Corrupt Good Math
CUSTOMER SEGMENTATION • Some Customers Behave Differently • Diversified Expertise • Perceive Your Values Differently
Histograms Present Numerical Data in a Way to Make a Point
o Snapshot of Data in Defined Groups o Just Enough Bars to Represent a Pattern
Identify the Distribution of Data by Shape
Clarity of Variability
Centermost Point
Visualize Data Trends
NEGATIVELY (LEFT) SKEWED POSITIVELY (RIGHT) SKEWED
Reading The Graph is Simple SYMMETRICAL
NON-SYMMETRICAL
Which Was Our Sample Data?
-2.83 -2.83 +2.83 +2.83 +2.83 -2.83
13, 18, 13, 14, 13, 16, 14, 21, 13
9.34 12.17 17.83 20.66 23.49 6.51
MEAN
15 POSITIVELY (RIGHT) SKEWED
A Nice Thing About A Histograms! BI-MODAL SPLIT
MULTI-MODAL SPLIT
Segmentation Errors The Impact of Improperly Segmenting Customers or Products
o Data Splits Create Averaging Errors o Leaving These Things Together:
Standard Deviation is Overstated
Mean Value is Wrong
Highs Are Too High
Lows Are Too Low
Drives Pricing Beyond Value Proposition
Overrides Will Reoccur Or Never Leave
EXCEPTIONAL ORDERS • Skews Data Positive or Negative
• Is Not A Typical Purchase • Should Be Excluded
When The Problem Is Not A Segment Dealing With Bad Behavior
SALES OUTLIERS • When The Segment Is Right • Performance Is Very Different • Best Profit Play Is Reclassification
THE EASY STUFF •Cost Overrides • Sell Contracts • Large Orders or Jobs
Getting Your Arms Around What Is Normal Exceptional Orders
NOT SO INTUITIVE • Sales Overrides
•Missing Contract Items • Special Pricing Agreements
EASY
HARD
o The Source File For Logical Analysis o The Evaluation Process
Typical System Sales
Exclude Obvious Deviations
Investigate Items Out Of Statistical Norms
Include and Exclude Data And View Changes
Prioritize Effort to Data With Greater Deviations and Dollars
Used To Derive The Analysis The Sales History File
Sales Outliers The Most Profitable And Accurate Matrix Allows Flexibility
More Like Market-Segment Editing
More Flexible Pricing Solution
o One Solution Is To Remove Them From The Calculations
o But Consider Movement Of Pricing To Align With Customer
Sales Outliers The Most Profitable And Accurate Matrix Allows Flexibility
More Like Market-Segment Editing
More Flexible Pricing Solution
Handles Exceptional Behavior
Maximize Profit, Minimize Risk
Prevents Averaging of Data
o One Solution Is To Remove Them From The Calculations
o But Consider Movement Of Pricing To Align With Customer
I Know What Happened, But Here Is What I WANT To Happen Blending Art With Science
Analytic Result
What The Limited Sales Data
Represents As Historically Accurate
Strategic Objective
The Financial Or Strategic Objective
Desired
Blending Art With Science Analytics = Where You Were; Strategy = Where You’re Headed
o Consider Customer And Vendor Relationships
o Limit Sales History To Reflect The Desired Commitment
New Vendor
Strategic Partner
New Products
Desired Customer
Unacceptable Performance
Timing Is Everything Vendor Price Increases Are The Perfect Storm
o Limited Sales Transactions
o Remove The Obvious
o Research The Not-So-Obvious
o Look For Patterns
o Quantify Overrides To Standard Deviation
o Eliminate Segment Errors!
o Execute Edits With Vendor Price Changes
Building A Price Matrix SUMMARY