revenue management: using data to drive price determinationsmmbc).pdf · revenue management: using...
TRANSCRIPT
Revenue Management: Using Data to Drive Price Determinations
Xin Chen
University of Illinois at Urbana-Champaign
Midwest Manufacturing Business ConferenceApril 25, 2017
Xin Chen
• PhD in Operations Research, MIT• Professor of Industrial Engineering at
the University of Illinois
• Research interest• Logistics and supply chain• Revenue management• Operations research
• Coauthor of “The logic of logistics”• Consulting: Anheuser-Busch,
Boxed.com, Jingdong.com, etc.
Operations Research (OR) and Analytics
OR is a discipline of developing and applying advanced data-driven analytical methods to help make better decisions
OR is at the core of Analytics 3.0
Fundamental Decisions Selling Products or Services
• When to list?
• What’s the asking price?
• Which offer to accept?
• When and how much to lower the listing price?
Fundamental Decisions Selling Products or Services
How to segment customers?
How to design products to prevent cannibalization across segments and channels?
What prices to charge in each segment, in different channels and over time?
Pricing and capacity allocations for different segments, channels and complements/substitutes?
Revenue Management (RM)
“Allocating the right type of capacity to the right kind of customer at the right price so as to maximize revenue or yield”
FOR EXAMPLE…
McGill, J. and G. van Ryzin (1999), Revenue Management: Research Overview and Prospects. Transportation Science, 33, 2, pp. 233-256.
Revenue Management
Revenue Management is concerned with demand-management decisions with the objective of increasing revenues and the methodology and systems required to make them.
RM refers to the wide range of Techniques Decisions Methods Processes Technologies
involved in demand management
Application Areas
Traditional Airline Hotel Extended Stay
Hotel Car Rental Rail Tour Operators Cargo Cruise
Non-Traditional Energy Broadcast Healthcare Manufacturing Apparel Restaurants Golf Sharing economy Advertisements More…
Dynamic Pricing is Standard Practice across Many Industries
Automobile
Apparel
PC
Airline Ticket
Time
Robert Phillips, Manugistics (now Uber)Source.…
Online Retailing
Forms of Dynamic Pricing
Personalized pricing Markdowns Display and trade promotions Coupons Discounts Clearance sales Auctions and price negotiations Surge pricing (Uber)
Conceptual Framework for RM Multidimensional nature
of demand Products, Customers, Time Locations, Channels, …
These dimensions are not independent Joint production capacity or
costs Customer behavior Information
©Talluri and Van Ryzin. Theory and Practice of Revenue Management
Sales - S Fixed Cost - F Variable Cost - c Price – P
Fc)S(PFScSP
−−=××=
)(Cost Fixed-)(Cost Variable -)( Revenue Profit
Price-based RM: Four Profit Levers
Common Ad-hoc Approaches to Set Prices
Cost-plus pricing
Competition-based pricing
Consumer-based pricing
Cost-Plus Pricing
Majority of businesses use cost-plus pricing.
How it works?Step 1: Determine sales target (e.g., Apple’s
iPhone: 5 million units)Step 2: Figure out the average cost to meet target
(e.g., $300 per unit)Step 3: Determine markup rate (e.g., 70%)Step 4: Price=Average Cost + Markup (e.g., $300+
$210=$510)
What is the margin of Ipad?
?%92260
260499Markup
=−
Why is Cost-Plus Pricing so Appealing?
Simple Price = Cost + Markup An entirely inward-focused exercise that has nothing
to do with the market.
Fair Merchants make only a fair living.
Financially Prudent Guaranteed profitability
Really? Any flaw in this logic.
Is Cost-Plus Pricing Fair
Firms have little incentive to reduce cost. (think about health care cost)
How much markup is fair for drug companies, giving the fact the production cost of a drug is close to zero? Or Ebooks?
Is Cost-Plus Pricing Always Profitable?
Demand is often uncertain
Cost also depends on the demand – nonlinear relationship
Cost itself could be uncertain (e.g., commodity price)
The Lessons of Cost-Plus Pricing
It is an inward-looking approach that distracts a company from its customer orientation and obscure the importance of market research.
It encourages ad hoc pricing decisions and bypasses many opportunities to improve pricing performances.
Competition-Based Pricing
Second most popular pricing approach
How it works?Step 1: Check out competitor’s priceStep 2: Adjust the price at about the same level,
mark up or down a bit
Easy and safe, isn’t it? No need to conduct any market research Setting a price close to competition price does
not risk losing market share Just follow “the invisible hand” to set the price
Boeing vs Airbus: The Price War In the mid- and late 1990s, Airbus was consistently gaining
market share and had surpassed its targeted “survival threshold” of 30% of market share.
Boeing decided to “beat back Airbus” and guard its 60% market share.
They were “making every bid a battle ground” and each would slash its price by at least 20% off the list price to grab an order. In 1995, Boeing lowered its price for 737s down from the list
of $35M to $19M, below its bottom line of $22M. Outcome: Losses all around!
Boeing guarded its market share with the cost of lowering its profit margin from 10% to 1% - a lower margin than a corner grocery store!
The Risk of Competition-Based Pricing
Lull the price setter into passivity
Managers lose sight of their own pricing responsibilities
Price-matching leads to a game of chicken –price war
Consumer-Based Pricing
Third common pricing approach
How it works Step 1: learn each customer’s willingness-to-pay
(e.g., car dealers) Step 2: charge different prices to different
customers (price discrimination)
Advantage of Consumer-Based Pricing
Flexibility to charge different prices to difference customers - set the price to match the size of customer’s wallet
Risk of Consumer-Based Pricing
Encourage comparison shopping
Focus on transactions instead of building a relationship and to channel creative energy into devising ways to win more or less money instead of forging a long-term, win-win partnership.
Risk of Consumer-Based Pricing
Often short sighted, alienate best customers with negative long-term consequences.
Over time, train the nice customers to become aggressive bargainers.
Risk of Consumer-Based Pricing
Demand more volatile(?)
Ignore operations reality Marketing and operations make
separate decisions Limited production capacity Available inventory
Here Comes Operations Research
Data Transaction Operations
Models Objective Constraints
Decisions Prices Production
Retailing Markdown Pricing
5%-15% gross margins improvement by the usage of model-based pricing software (Friend & Walker 2001)
Style-GoodsMarkdown
Pricing
ConsumerPackage-goods
Promotion
• Apparel, sporting goods, high-tech andperishable foods
• High initial price, markdown low reservation items
• Markdown on peak periods: more sensitivecustomer behavior
• Soap, diapers, coffee, yogurt, …• Customer awareness of past prices/promotions• “reference price”• “stockpile” behavior
(c)2005 Robert G. Cross
Revenue Management-Process Flow
Customer Demand Data Forecasters
Optimization
Price and Availability Recommendations
MarketSegmentation
Constraints
?
$Market!
(c)2005 Robert G. Cross
How Revenue Management Optimizes ProfitFor Any Business
PredictCustomerDemand
OptimizePrice
RecalibrateDynamically
Segmentthe Market
Segmentation based on buying behavior, not just current or past classifications
Forecasts of demand and capacity at product/price level
Mathematically determine capacity availability and price that maximizes expected profit
Continually monitor performance and update market response
(c)2005 Robert G. Cross
Case StudyFord Motor Company
The Problem: How to Optimize Customer Incentives over millions of vehicles
The Solution:
• Geography
• Vehicle Type
• Incentive Device
• Competition
• Response to PromotionBy Segment
• Probability of “Winning”
• Competitive Response
PredictCustomerBehavior
OptimizeResponse
DynamicallyRecalibrate
Segmentthe Market
• Vary Discount
• Vary Device
• Vary Time
• Measure Response
• Update Models
Result: Expected savings of $800 million per year
(c)2005 Robert G. Cross
Case StudyUnited Parcel Service
The Problem: How to Optimally Price a Hundred Thousand Annual Contracts
The Solution:
• Region
• Customer Type
• Competition
• Strategic Goals
• Probability of “Winning”
• Probable Margin
• Calculate Optimal Point Maximizing Probability of Win at Greatest Margin
• Measure Success
• Update Models
PredictCustomerBehavior
OptimizeResponse
DynamicallyRecalibrate
Segmentthe Market
Result: Increased revenue and profit in excess of $120 million per year
Coca-Cola Targets Small Retailers With Coke on Demand App (Wall Street Journal)
• Coca Cola Co. is encouraging small mom-and-pop retailers to use a mobile app that notifies the soda giant when the retailers need new product.
• Businesses use the app to send their inventory needs to local distributors, which bid on the deal.
• The distributors agree to a certain price and time period for delivery. Orders are filled in between 45 minutes to four hours, which compares to two or three days by traditional trucking.
Demand Modeling under Dynamic Pricing
• A predominate assumption is that demand depends on current price only
• For frequently purchased consumer goods, consumers’ purchase decisions are heavily influenced by past observed prices
• Reference price: A price consumers formed in mind against which they compare with shelf price
Dynamic Demand Models: reference price
Consumers’ behavior Form a reference price 𝑟𝑟: 𝑟𝑟+ = 𝛼𝛼𝑟𝑟 + 1 − 𝛼𝛼 𝑝𝑝 Compare to the current price 𝑟𝑟 − 𝑝𝑝 Decide the purchase depending on whether
Gains if 𝑟𝑟𝑡𝑡 > 𝑝𝑝𝑡𝑡 Losses if 𝑟𝑟𝑡𝑡 < 𝑝𝑝𝑡𝑡
Aggregate level demand 𝑑𝑑 𝑝𝑝, 𝑟𝑟 = 𝑏𝑏 − 𝑎𝑎𝑝𝑝 + 𝜂𝜂(𝑟𝑟 − 𝑝𝑝)
Aggregate Model in Practice
• Canned Tuan data from Dominick Finer Foods database maintained by Chicago Booth
• Greenleaf (1995): Reference price model provides a much better fit to the data in peanut butter category.
• Raman and Bass (2002): Four out of five brands show significant Evidence.
Gain-Seeking Special Case: Dynamics
When consumers’ initial reference price is low, use a regular price.
Then apply a skimming pricing strategy by gradually discounting the regular price overtime to drop the consumers’ reference price.
Business Conditions Conducive to RM
Production inflexibility Customer heterogeneity Demand variability and uncertainty Data and information systems
infrastructure Management culture