Chapter 3

Question 1

Explain: What is the negative binomial distribution model/

Answer 1

BDM: It captures the weight (EG how many people will by once, twice, three times etc) and timing (when they will make that purchase) in the specific buying category.

Question 2

What does the negative binomial model do?

Answer 2

Explains and predicts consumers buying of categories

Question 3

What are the two parts of the NDB model?

Answer 3

• Gamma Distribution: Captures the purchase weight

- Poisson distribution: Captures the purchase timing

Question 4

What is Penetration?

Answer 4

Symbol: b

It is the proportion of the market that bought the brand/category at least once in the given period

Question 5

What is Average Purchase Frequency?

Answer 5

The average number of times a buyer the brand buys in a specific period

Question 6

Are markets stationary in the medium term?

Answer 6

• Yes

- Markets are generally stationary in the medium term because buyers are habitual

Question 7

How do you describe how people buy in a given period?`

Answer 7

It occurs as-if-randomly

Question 8

Describe the zero order process

Answer 8

Buyer behaviour is a zero order process
What was bought in the previous transition has no increasing or decreasing chance of making the consumer by that particular brand or category next time
- It is a probability thing, if you bought a product 80% of the time, your likely to buy it again

Question 9

Are buyers generally light buyers?

Answer 9

• Yes
• There are very few heavy buyers
• Follows the paterto rule but not in the 80-20 proportions that are proposed
• We assume that it is likely 50% of category volume comes from the top 20% of buyers
• Light buyers and non-buyers are very important

Question 10

Do bigger brands have more buyers?

Answer 10

• Bigger brands have more slightly heavy buyers

- These bikers account for a smaller portion of the brands total sales than for a smaller brand

Question 11

How often are categories bought?

Answer 11

• Most categories are only bought once on each purchase occasion
• A brands customers buy competing brands more often then they buy one particular brand
• Repeat buying is normal 50% (50% of those who bought last time will buy again next time)

Question 12

Over time are light buyers important?

Answer 12

Over longer periods a brand has more light buyers who are worth more to the brand.

Question 13

How are purchases stationary in the medium term?

Answer 13

• At the category level there is stationarity in the size of the market. Most markets have been purchased by people for some time. The numbers of shoppers that purchase from a category does no dramatically change.
• At a brand level: meaning that market share doesn’t change that much, in the medium term. This doesn’t mean that there are not short -lived changes in market share, as these can be quite significant for a short period as a result of a price promotion, for example. Nor does it mean that trends do not occur over the longer term (i.e. Holden’s rise in popularity over Ford over the past 10-15 years).

Question 14

How do we know most markets are stationary for many years?

Answer 14

• Empirical Observation
• Dirichlet Model:The fact that the Dirichlet Model fits most markets pretty well means that empirically this assumption is correct i.e. it means that markets are actually stationary, if they weren’t the model would not fit the data, and this is indeed what we see when markets change their market share (a lack of model fit).

Question 15

Why does it matter that markets are stationary?

Answer 15

• Suggests marketing activity does little to influence buyers in the medium term
• Marketing in the short term has a better impact

Question 16

What is Gamma Distribution?

Answer 16

The Gamma distribution describes the number of purchases of the category that people make within a time period.

Question 17

How does a gamma distribution show the importance of light buyers?

Answer 17

• It shows that there are many light buyers (and non-buyers) and very few heavy buyers.
  -As you measure longer periods of purchasing behaviour, you will find more
  and more non-buyers become light buyers by making purchases and essentially coming into the market.

Question 18

20-80 rule

Answer 18

The top 20% of a brand’s customers (heavy buyers) generally account for 50% of sales and the remaining 80% of lighter buyers also make up 50% of sales. So light buyers matter because there are lots of them. As a group they account for a large proportion of a brand’s sales, just because of their sheer number. Many of any brand’s customers are light buyers of the brand because they also buy from other brands in the category (i.e. repertoire buying).

Question 19

Draw a gamma distribution graph

Answer 19

As number of times/purchases increase, number of buyers/households decrease

Question 20

What is the first order effect?

Answer 20

ach purchase that a consumer makes has some impact on later purchases

Question 21

Is the first or zero order assumption more reasonable?

Answer 21

The Dirichlet (along with several other stochastic models of buyer behaviour) has a zero order assumption. Empirically we find that the zero order assumption is most predominant in the majority of markets (i.e. for most purchases, past purchase of a brand does not make someone more likely to purchase that same brand next time), therefore, this is what we need to follow.