jump to navigation

Value of Retailing Services June 12, 2007

Posted by jyu in Qualifying.
add a comment

–Manufacturers sell a certain good for which consumers have demands

–Retailers sell services in addition to goods.  Retailers sell to final consumers what someone else has made through phsical store or internet.  Retailers exist and operate profitably because they perform a service which their customers value.  The are then compensated through margins by marking-up goods they sell.

According to Bliss (1988), the retailer incurs overhead costs which must be recovered by markups.  The consumers also incur overhead costs in coming to the shop and would incur additional costs if they have to visit more than one shop.  An obvious condition of equilibrium is that all shops should offer equally good vlaue for moeny as measured by the indirect utility function common to all consumers.  A shop offering less good value than others would lose customer, while a shop offering an excess of value could raise some prices and increase its profit.  This is the reason why retailers provide various services along with the good that they sell to consumers.

Services provided by retailers include:

  • advertising
  • labor staffing
  • one-stop shopping
  • assortment
  • convenient location…

Most of these services improve information or cut down on search and transaction costs. Ratchford and Stoops (1992) illustrated that marginal contribution of retail advertising and labor as the marginal reduction in the consumer’s acquisition costs.  They key concept is that retail services will be traded off for increased margins up to the point where marginal saving for consumers is equal to marginal increase in the retailer’s margin.

Although manufacturer might also provide advertising to influence consumer preference, the advertising primarily serves to emphasize product differentiation and hance mostly increases the full price paid by consumers.

Retail Formats:

  • Bentancourt and Gautschi (1990): provide a formal analysis of the retail assortment problem by formulating the retail demand as a function of the underlying demand for various activities involved in the household production.  Based on this view, a household pursing a disaggregated consumption activity like dinner on Christmas eve may choose to shop at a specialty store with a greater depth of assortment than at a convenience store.  Thus, different retail formats exist to satisfy varying demand of the underlying consumption activities
  • Messinger and Narasimhan (1997): develop a model to explain the growth of one-stop shopping and suggests that greater prevalence of one-stop shoping has been a response to growing demand for time-saving convenience

Managing Channel Profits (Jeuland & Shugan 1983) June 12, 2007

Posted by jyu in Qualifying.
add a comment


  1. Channel coordination problems occur with all marketing decision variables albeit in different directions.  Without coordination, marketing effort will be smaller than optimum.  This is a generalization of the result concerning margins: without coordination, they will be larger than optimum
  2. Achieving channel coordination can be difficult.  However, several mechanisms do exist for achieving coordination (e.g.  joint ownership, Simple contracts, implicit understand, profit sharing, and quantity discounts)
  3. Many channel phenomena (e.g. integration, contracts) may be implicit coordinating mechanisms
  4. Joint ownership and fixed price contracts are often inadequate mechanisms for coordination
  5. Quantity discounts can provide an optimal means for achieving coordination
  6. Quantity discounts can take the form of other marketing phenomena such as cooperative advertising or added service levels
  7. Quantity discounts are a method of profit sharing
  8. The channel coordination can be separated from the profit division issue.  Although, they are related decisions.
  9. A coordinated channel will make R&M’s margins appear to be too low
  10. Coordination, once achieved, will lead to lower margins, higher levels of marketing effort, lower retail prices and larger total channel profits

Signaling and Screening June 12, 2007

Posted by jyu in Qualifying.
add a comment

Two types of information asymmetry:

  1. Hidden action (imperfect information) – Principal tries to induce the agent to take the single appropriate action
  2. Hidden type (incomplete information) – Principal tries to induce the agent to take the action that will reveal his type

Moral Hazard (in the context of hidden actions):

–A form of post-contractual opportunism that arises because actions that have efficiency consequences are not freely observable and so the person taking them may chose to pursue his or her private interests at other’s expenses

Adverse Selection (in the context of hidden types):

–Asymetric information about the pre-contractual information relevant to the transaction between the principal and agent

–The tendency of those in dangerous jobs or high risk lifestyles to get life insurance or seller has information about product quality but buyers don’t

Solutions to asymmetry information:

  1. Hidden actions: The principal develops incentive contracts (Mechanism)
  2. Hidden types: the informed party can use signaling while the uninformed party can use screening

Signaling: Agent conveys some meaningful information about himself to principal (signal before principal offers a contract)

Screening: Principal moves first or suggest a contract and learn agent’s type through his behavior (signal after a contract)


  1. Manufacturer – Retailer: A manufacturer charges higher price or advertise more to convince that his product is in high quality (signaling).  A retailer asks “slotting allowance” to find the type of the product – high/low demand (screening)
  2. Used Car Trade: Seller proposed warranty/build reputation (signaling).  Buyer do a mechanic check on the car (screening)
  3. Education in the labor market: Employee obtains education to show that he is high performance type (signaling).  Employer chooses only graduate students expecting higher performance (screening)

Optimal Search Strategy June 12, 2007

Posted by jyu in Qualifying.

Choice                    A                          B

Probability      0.1      0.9           0.3       0.7

Payoff             $100    $0            $6        $4

Cost                        $5                         $2

The expected payoffs of the two alternatives, without search, are

A: -5+100*0.1+0*0.9 = 5

B: -2+6*0.3+4*0.7=2.6

Supposed we search B first and the A second:

If we see payoff $6 first, then the payoff is -5+(100*0.1+6*0.9)=10.4

If we see payoff $4 first, then the payoff is -5+(100*0.1+4*0.9)=8.6

The expected payoff of this sequential search is

-2+{[-5+(100*0.1+6*0.9)]*0.3+[-5+(100*0.1+4*0.9)]*0.7} =7.14

Suppose we search A first and then B second:

If we see payoff $100 first, then the payoff must be $100 since it is the highest it can get

If we see payoff $0 first, then the payoff is -2+(6*0.3+4*0.7)=2.6

The expected payoff of this sequential search is


Apparently, consumer should search A first and then B second.  The expected payoff of this sequential search is the highest.

EG about market evolution and stationarity June 12, 2007

Posted by jyu in Qualifying.
add a comment

Evolution is the dominant characteristic for sales and marketing mix spending, but that stationarity is the dominant characteristic for market share.


  1. Sales evolution is more likely to occur at the industry/category level than at the individual brand/firm level
  2. Sales and market share evolution in durables is as likely as in non-durables
  3. Temporal aggregation affects the likelihood of finding evolution
  4. Sales evolution was most freqeuntly observed in the seventies
  5. The longer the sample, the more likely one is to find evolution in sales, but not in market share
  6. There is a significant difference between North America nad Europe in their porportion of evolving sales variables

Promotions are mixed strategies June 12, 2007

Posted by jyu in Qualifying.
add a comment


  1. Competitive promotions are mixed strategies
  2. Depth of promotion has a bimodal distribution (on more limited data)

Competitive promotions are mixed strategies:

  • First an empirical regularity is established that promotions are independent across competitors.
  • This regularity is then elaborated on in the context of a promotion game
  • The promotion game is linked to observable outcomes, and a classification of possible situations is developed.  In particular, the classification includes the prisoners’ dilemma, battle of the sexes, and marketing models of promotion competition
  • The evidence for the generalization comes from a variety of product markets, spanning trade promotions, retail price reductions and retail promotions such as advertised specials.  The product markets include coffee, baby diapers, toilet tissue, saltines, dishwashing fluid, ketchup and detergents among others

EG of the effects of deals and promotions June 12, 2007

Posted by jyu in Qualifying.
add a comment
  1. Temperal retail price reductions substantially increase sales
  2. Higher market share brands are less deal elastic
  3. The frequency of deals change the consumer’s reference price (if a product is promoted heavily – discounted deeply and promoted frequently – the consumer’s reference price of the product decreases.  The consumer will then buy less of the product at regular price because his or her reservation price has decreased correspondently)
  4. The greater the frequency of deals, the lower the height of the deal spike
  5. Cross-promotional effects are asymmetric, and promoting higher quality brands impacts weaker brands (and price label products) disproportionately. (brand 1 can capture significant share from brand 2 – hence the asymmetry – since brand 1 can use promotions more effectively than brand 2.  Brand 2 cannot easily retaliate because of the asymmetry in promotional response.)
  6. Retailers pass-through less than 100% of trade deals (some portion of funds spent by manufacturers to stimulate retailer promotion is pocketed by the retailer to enhance their profits.)
  7. Display and feature advertising have strong effects on item sales
  8. Advertised promotions can result in increased store traffic
  9. Promotions affect sales in complementary and competitive categories

Science and Empirical Generalizations June 12, 2007

Posted by jyu in Qualifying.
add a comment

EG is a pattern or regularity that repeats over differnt circumstances and that can be described simply by mathematical, graphic or symbolic methods.  A pattern that repeats but need not be universal over all circumstances.

Science is a process in which data and theory interact to produce higher level explanations.  It consists of (1) empirical generalizations and (2) generalized explanations of EGs.

Marketing science is a process that involves EGs, generalized explanations, as well as testing and revision or extension of the generalized explanation.

A higher level explanantion will reduce as a special case to a lower level explanation, and it will possibly suggest the existence of additional phenomena.

Example:  Bass Model is an EG of the difussion of innovations.  GBM is a higher level theory.

MCMC June 12, 2007

Posted by jyu in Qualifying.
add a comment
  1. MCMC methods are ideally suited for models build from a sequence of condotional distribution, often called hierarchical models.  Bayesian hierarchical models offer tremendous flexibility and modularity and are particularly useful for marketing problems
  2. Marketing applications using Bayesian methods include: Discrete Choice Models, Conjoint Analysis, Effects of purchase timing, satiation, determinants of heterogeneity, brand preferences, and etc.
  3. The emergence of MCMC methods has eliminated the computation bottleneck related to the Bayes Theorem.  MCMC substitute a set of repetitive calculations, that in effect, simulate draws from the posterior distribution.  These MCMC draws are then used to calculate statistics of interest such as parameter estimates and confidence interval.  The idea behind MCMC engine that drives the HB revolution is to set up a Markov chain that generates draws from the posterior distribution of the model parameters
  4. To obtain posterior results, MCMC simulation is often used.  The unobserved vairables may be siulated alongside the model parameters from their posterior distribution.  This technique is called Data Augmentation.  Given simulated unobservables, obtain the likelihood function conditional on the unobservable variables.
  5. Two often applied MCMC samplers are the Gibbs Sampler and the Metropolis-Hasting sample.
  6. The Gibbs sampler is a MCMC method of sampling probability densities.  This method has made possible the Bayesian approach to the estimation of nonlinear panel data models providing accurate finite sample estimates
  7. Gibbs Sampling: an algorithm to generate a sequence of samples from the joint probability distribution of two or more random variables.  The purpose of such a sequence is to approximate the joint distribution or to compute an integral.  Gibbs sampling is a special case of the Metropolis-Hasting.  It is applicable when the joint distribution in not known explicitly, but the conditional distribution of each variable is known.
  8. Metropolis-Hastings algorithm is rejection sampling algorithm used to generate a sequence of samples from a probability distribution that is difficult to directly sample from.  This sequence can be used in MCMC simulation to approximate the distribution or to compute an integral. 

Steps of algorithm implemented for estimation of the random effects Tobit model:

  1. Run a GLS estimation with the original truncated data to fix the initial values
  2. Sample the censored variables to build the argumented dataset
  3. Run a GLS estimation on the panel with the augmented dataset for computing new mean values
  4. Draw Betas from the distribution
  5. Estimate the individual effects using the residuals from the previous steps
  6. Draw w from the distribution
  7. Draw sigma from the distribution

Game Theory Definitions June 11, 2007

Posted by jyu in Qualifying.
add a comment

Nash Equilibrium (NE) (after John Nash):

A kind of solution concept of a game involving two or more players, where no player has anything to gain by changing only his or her own strategy unilaterally.  In other words, if no player has incentive to deviate from his strategy given that the other players do not deviate.

Subgame Perfect Equilibrium (SPE): 

  1. a refinement of a NE used in dynamic games
  2. a strategy set is a SPE if it represents a NE of every subgame of the original game
  3. the players palyed at any smaller game that consisted of only one part of the larger game and their behavior represents a NE of that smaller game, then their behavior is a SPE of the larger game
  4. Common method for determining SPE is backward induction

Perfect Bayesian Equilibrium (PBE):

A PBE is a strategy combination and a set of beliefs such that a strategy combination consisting of the best responses given that equilibrium beliefs follow Bayes’s rule and out-of-equalibrium beliefs follow a specific pattern that does not contradict Bayes’s rule. 

Folk Theorem: 

A class of theorems which imply that in repeated games, any outcome is a feasible solution concept, if under that outcome the players’ minimax conditions are satisfied.

The minimax condition states that a player will minimize the maximum possible loss which they could face in the game

Pure Strategy:

A pure strategy maps each of a player’s possible information sets to one action.

Mixed Strategy:

A mixed strategy maps each of a player’s possible information sets to a probability distribution over actions.

Incentive Compactibility Constraint (IC):

The IC constraint takes account of the fact that the agents move second, so the contract must induce him to voluntarily pick the desired effort

Individual Rationality (participation constraint) IR:

The agents prefer the contract to his reservation utility.

Incomplete Information:

The utility payoffs of each player remain private ifnormation of each player.  Games with incomplete information require the players form beliefs about their opponents’ private information

Minimax Punishment:

Minmax strategy is defined as the most severe sanction possible if the offender does not cooperate in his own punishment.  The set of strategies is a set of minimax strategies chosen by all the players except i to keep i’s payoff as low as possible, no matter how he responds