Notes for Marketing

Session 1 Marketing

Understand (Needs & wants of) customers and create value -> Growth.

1. Structure

Market Analysis: Content, Company, Customers, Competitors, Collaborators

Strategy: Segmentation, Targeting, Positioning

Mix: Product (create value), Promotion (communicate value), Place (deliver value), Price (capture value)

2. Value

Value = Perceived Benefits - Perceived Costs

• Perceived Benefits: Rational(economic), Functional, Psychological(emotional, social)

To influence experience:

• Bottom-up processes: can be sensed
• Top-down processes: Expectations, desires, beliefs

3. Best buy case

Customer Decision Process:

• Need Recognition
• Evaluation -> Best Buy (provides value to brands) / Circuit City
• Choice -> Best Buy
• Purchase -> Best Buy replaced by Amazon

Value is created along the process.

Customers can be anyone who gets value from you.

Response to Covid-19:

• curbside pick up
• by-appointment shopping
• stores as fulfillment center
• partner with Shipt, quicker delivery

*Case mentioned

Sell milkshakes - customer insight, motivations

Toyota Prius vs Porsche 911 Carrera

Tinder: what value to customers?

Coke Cola: emotional connection

SoBe Adrenaline Rush: How to increase mental acuity

Session 2 Understand the Market

1. Causation

Online Advertisement

Q: eBay Keyword ads(SEM, Search Engine Marketing) -> sales?

Fact: Customers who saw the ads are supposed to have higher sales

Confounding variable: Familiarity with eBay

Conclusion: Natural search results is almost perfect substitute results: 99.5% from paid -> natural.

Blake, T., Nosko, C., & Tadelis, S. (2015). Consumer heterogeneity and paid search effectiveness: A large‐scale field experiment. Econometrica, 83(1), 155-174.

Be aware of random correlations.

2. Experiments

1) Components

• Test question
• Independent Variable
• Randomization - Selection bias: not equal likelihood of being in each group
• Dependent Variable

2) Goals

Internal Validity

• Confounds

External Validity

• Sample -> population (generalization)
• Context

3) Advanced Designs and Analysis

Full factorial design

• change two independent variables

A/B Test

• webpage design: media * buttons = combinations cells

4) Intent to treat vs Treatment on treated

Effect of program? Effect of treatment?

• Email openers may be bigger spenders.

Promotion(ITT), claimer/redeemer(TOT)

3. Sampling

1) Random

2) Matched samples: more efficient

3) Sample size?

Statistical power

• in R: power.t.test(n=NULL,delta=7,sd=15,power=0.90,sig.level=0.01) gives n=138.3163

dependes on:

• delta (effect size/difference): What’s the smallest you would care about?
• Sample size: What’s the largest sample you could get?

4. Quanlitative Methods (for less well-defined questions)

1) Focus Groups & Interviews

2) Observation (do not intercept)

5. Surveys

error:

• Sampling error (too few samples, not representative)
• non-sampling error
  • Non response error (selective non response, e.g. Hotel satisfaction,AT&T bills)
  • Measurement error
    • Question ambiguity
    • Order effects
    • Question fails to capture construct
    • Social desirability

Coping with non-random samples

• Validate large(cheap) non-random by comparing with samll(expensive) more random
• weights/quotas: make sample size $\times$ weight equal for all groups. (weight = Target(customer base) proportion / Survey propotion)

More about Weighting:

• Pew Survey Analysis on Weighting
• Qualtrics Information on Weighting

*Case mentioned

Hormone replacement therapy: Women with higher socioeconomic status got the new therapy in study

Ice-cream sold vs Murder rate

Divorce rate in Maine vs Per capita consumption of margarine

people drown in pool vs films Nicolas Cage appeared in

Gym attendance experiment

Random digit dialing - Survey

Session 3 Customer Lifetime Value

1. Interview with Mitch Lowe, CEO of MoviePass

MoviePass:

Is is profitable?

But customers may unsubscribe (Churn), casual moviegoers will cancel at a greater rate.

2. Customer Lifetime Value

Given margin earned at the end the period 1,

$$ CLV = m(\frac{r}{1+i-r})-AC $$

where m is margin per period, i is discount rate, r is retention rate = 1-churn rate, AC is acquisition rate.

3. Increasing CLV

increase r

decrease AC

• Value of customers differ across channels. Identify high value customers

increase margin

• up-selling, cross-selling

*Case mentioned

Hubspot

BlueApron

Groupon

Session 4 Harvest value: Pricing

1. Pricing methods

1) Cost-driven

2) Competition-driven

3) Value to Customer

Value to Customer = Reference Value + Differentiation Value

• Reference Value: Price of Closest Competitive offering (CCO) / best substitute
• Differentiation Value

Economic Value to the customer (EVC) is the maximum amount a customer should be willing to pay assuming fully informed about benefits

Perceived value (WTP) is lower? Increase the expectation

Differentiation value can be negative. e.g. RYANAIR Airlines

2. What values?

Conjoint: Valuing individual attributes

Learn more from Sawtooth Software

Ask for relative comparisons of different products with price as an attribute.

Run regression and the coefficients are utilities, or “part worths”

Usage of conjoint analysis:

• Segmentation
• Product development
• Predict choice
• Pricing
• Brand equity

3. Discrimination

Identify differences in WTP based on

1. Occasion

• for wedding, etc

2. Timing

• during peak times

3. Quatity (multi-part tariff)

• Customers can become more price sensitive with higher quantities purchased

4. Customer characteristics

• Early adopters have higher WTP

4. Consumer Psychology

Prospect Theory: pain from loss is more

• discount is gain

5. Calyx Flowers Case

Both are upscale professional women

Cost per catalog: ~$0.75

Average margin: ~$40

Business Model:

• Calyx:

Competitive Analysis:

Segmentation:

• Demographics
  • Original: Age 30-65, income > 100k, female
  • Possible: Age 30-65, income > 100k, male; Age 30-65, income 50k-100k, female
• Needs/Occasions:
  • Gifts
  • Events
  • Business/decoration
  • Home decoration/Hobby

Needs-based segmentation - More efficient targeting

*Case mentioned

Atlantic Computers

Steves Jobs Offers Rare Apology, Credit for iPhone

Economist Subscription Pricing

Session 5: STP

1. Segmentation

Never try to please everybody

Base of Segmentation

Cluster Analysis - How many clusters?

Statistical:

• Elbow Method
• Silhouette Method

Managerial:

• Capability of company
• Sufficient size of cluster
• Sufficient, identifiable differentiation

2. Targeting

Customer (Opportunity for Profit)

• Segment Size
• Growth Rate
• Reach ability & ability to fulfill needs

Company (Fit)

• With Objectives
• With Competencies
• With Resources

Competitive (Intensity)

• Underserved needs
• Competitors’ strengths

3. Positioning (in the minds of consumers!)

• Target Market
• Key differentiator
• Competitive set
• Evidence you claim

Perceptual Maps

• Factor analysis

Positioning • What brands are we seen as similar to? • What associations to customers have with our brand (and our competitors’ brands)? Competitive landscape • Who are our closest competitors • Which customers are we competing for Market Opportunity • Are there any unmet needs in the market?

4. Mini case

Background: Leader in the B2B market for pregnancy test

Objective: Wants to gain share in the B2C market

STP + 4Ps

*Case mentioned

Hillrom

Planet Fitness

Best buy: firing customers

LinkedIn vs other social media

VOLVO: Safe

Vij’s restaurant

Cadillac: from a “classic American” image to “young luxury” image