[PPT] - The Importance of Rank Or How Our Brains Constrain Survey Responses PowerPoint Presentation

SLIDE 1

The Importance of Rank

Or How Our Brains Constrain Survey Responses Or The Enormous Power of Winning

Kyle Findlay Senior R&D Executive The TNS Global Brand Equity Centre

SLIDE 2

SLIDE 3

The enormous power of winning

This presentation is about how these inequalities emerge

1 2 3 4 5 1 2 3 4 5

SLIDE 4

“

One is wonderful.

Two is terrific.

Three is threatened.

Four is fatal ”

~ Larry Light

SLIDE 5 Retailers (UK) Retailers (China) Auto (UK) Auto (Germany) Sources: Retailers = KWP, Auto = Mintel

SLIDE 6 Source: Kohli, R and Sah, R. 2003. “Market Shares: Some Power Law Results and Observations”, Harris School Working Paper, Series 04.1

“Market share versus rank across 506 food brands and 665 sporting-goods brands

“ ”

Power laws!

SLIDE 7

Man can be thought of as perfectly rational

John Maynard Keynes

SLIDE 8                        

SLIDE 9

The ‘Chicago School’ of market research            

SLIDE 10

Nope. Humans have bounded rationality

Herbert Simon

SLIDE 11

Markets aren’t fair

SLIDE 12

We make good enough choices from a ‘stacked deck’

Sources: http://www.pepsico.com/Annual-Reports/2008/performance/n-america-csd.html

SLIDE 13 Sources: http://www.datapointed.net/visualizations/maps/distance-to-nearest-mcdonalds

SLIDE 14 65.1 38.9 28.2 25.2 24.0 17.0 2.5 1.2 0.9 0.7 0.7 0.5

Newspaper 1

Newspaper 2

Newspaper 3 Newspaper 4 Newspaper 5 Newspaper 6 Usage of brand Frequency of purchase Source: UK newspaper market

SLIDE 15            

We do the best we can within our cognitive limits

SLIDE 16

Get the important bits right and the rest will follow



 



SLIDE 17

The brain is impressive…

SLIDE 18

The brain is impressive… but it does have limits

SLIDE 19

Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand

SLIDE 20

Brand Brand Brand Brand Brand

Used

Consider

Used

Consider

Used

Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand Brand

SLIDE 21

1,058,290

respondents

1,267

studies

88

countries

205

3.9

Mean

Standard deviation = 3.7 Data: 1,267 studies, 88 countries, 205 categories, 1,058,290 respondents

3.0 2.0

Median Mode

So, just how many brands do we realistically hold in mind at a time?

SLIDE 23 Standard deviation = 3.7 Data: 1,267 studies, 88 countries, 205 categories, 1,058,290 respondents 17.9 23.1 15.6 11.4 8.0 6.0 4.3 3.9 1.4 0.9 3.9 1 2 3 4 5 6 7 8 9 10 The rest % of observations Evoked set size

SLIDE 24

2.3

Auto

3.3

Finance

5.0

Consumer

3.1

Energy

3.3

Business Services

3.4

Technology

2.3

Polling & Social

3.1

Healthcare Standard deviation = 3.7 Data: 1,267 studies, 88 countries, 205 categories, 1,058,290 respondents

SLIDE 25

Beer

2.9

Thailand

8.5

Poland

10.1

UK

8.0

Spain

8.5

France

8.7

Germany

2.8

India

3.7

Cambodia

SLIDE 26 10 20 30 40 50 60 70 80 90 100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 # brands in brand list Average evoked set size

0.36 0.13

Correlation Linear R2

1,267 studies

Standard deviation = 3.7 Data: 1,267 studies, 88 countries, 205 categories, 1,058,290 respondents

SLIDE 27

Evolution

made us care about

SLIDE 28 Sources: http://www.supersport.com/olympics/gallery/22921

SLIDE 29 Sources: http://xkcd.com/1098

SLIDE 30 Source: https://www.checkmarket.com/2011/06/net-promoter-score

SLIDE 31 Source: Louw & Hofmeyr (2012) Reality Check: The Relationship Between What We Ask and What People Actually Do

SLIDE 32 69.5 23.6 11.3 4.1 R² = 0.97 1 2 3 4 Share of wallet (panel data) Share of wallet rank

Power law!

Rank 1 brands get most of the share

n=984 UK laundry detergent | Actual panel data supplied by KWP

SLIDE 33 UK laundry detergent | Actual panel data supplied by KWP 10% 20% 30% 40% 50% 60% 70% 80% 90% % of observations Share of wallet

Rank 1 Rank 2 Rank 3 Rank 4

SLIDE 34 x y logx logy

SLIDE 35

Power law!

Source: Krumme, C, et al (2013) The predictability of consumer visitation patterns Caption: “Probability of visiting a merchant, as a function of merchant visit rank, aggregated across all individuals. Dashed line correspond to power law fits P(r) , r2a to the initial part of the probability distribution with a 5 1.13 for the European and a 5 0.80 for the North American database”

Retailer visits by rank

Power law!

SLIDE 36 60.2 32.5 18.9 17.7 13.3 10.7 9.0 7.0 5.6 5.4 4.8 3.1 3.9 2.1 3.6 1.4 2.2 2.0 4.1 2.6 2.3 R² = 0.93 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11.0 Share of wallet (panel data) Performance rank

Power law!

Top rated (i.e. rank 1) brands get most of the share

UK laundry detergent | Actual panel data supplied by KWP

SLIDE 37

too much

information

It is possible to measure

SLIDE 38

First mention

i.e. rank 1

0.57

0.28

0.17

All unaided aware

i.e. rank 2, 3…

Aided aware

i.e. All ranks Awareness metric Countries: UK, China | Categories: Laundry, Retail Stores | Number of datasets: 5 | Actual panel data supplied by KWP Source: Hofmeyr, J & Louw, A. 2012. Reality Check : The Relationship Between What We Ask and What People Actually Do. ESOMAR 3D Conference 2012, Amsterdam Respondent-level correlation with spend

SLIDE 39 Respondent-level correlation with spend Usage metric

Most often

i.e. rank 1

0.71

0.58 Past 3 months

0.37 Ever bought i.e. all ranks Countries: UK, China | Categories: Laundry, Retail Stores | Number of datasets: 5 | Actual panel data supplied by KWP Source: Hofmeyr, J & Louw, A. 2012. Reality Check : The Relationship Between What We Ask and What People Actually Do. ESOMAR 3D Conference 2012, Amsterdam

0.62

0.56

Regularly

i.e. rank 2, 3…

Past 1 month

SLIDE 40 78 36 36 36 Cell 1 Cell 2 Cell 3 Cell 4

# attributes in survey

Less measurement; richer data

SLIDE 41 78.0 17.0 18.0 28.0 Cell 1 Cell 2 Cell 3 Cell 4

# attributes selected

SLIDE 42 2.2 3.0 3.2 4.8 Cell 1 Cell 2 Cell 3 Cell 4

# ticks per respondent

SLIDE 43 13.0 3.0 3.5 5.5 Cell 1 Cell 2 Cell 3 Cell 4

Total survey time

SLIDE 44

Conclusions

Humans have bounded rationality



 



SLIDE 45

3.9

SLIDE 46 1 2 3 4

SLIDE 47

First mention Most

ften

Less is more

SLIDE 48

Data Respondent Time Money

SLIDE 49

Thank you

Acknowledgements:

Anna Retief Elanie de Beer Jannie Hofmeyr Ken Bell Bruno Gonçalves Constantin Michael

SLIDE 50

First mention

i.e. rank 1

0.57

0.28

0.17

All unaided aware

i.e. rank 2, 3…

Aided aware

i.e. All ranks Awareness metric Countries: UK, China | Categories: Laundry, Retail Stores | Number of datasets: 5 | Actual panel data supplied by KWP Source: Hofmeyr, J & Louw, A. 2012. Reality Check : The Relationship Between What We Ask and What People Actually Do. ESOMAR 3D Conference 2012, Amsterdam Respondent-level correlation with spend

0.92

0.81

0.68

Aggregate-level correlation with spend

SLIDE 51 Respondent-level correlation with spend Usage metric

Most often

i.e. rank 1

0.71

0.58 Past 3 months 0.37 Ever bought i.e. all ranks Countries: UK, China | Categories: Laundry, Retail Stores | Number of datasets: 5 | Actual panel data supplied by KWP Source: Hofmeyr, J & Louw, A. 2012. Reality Check : The Relationship Between What We Ask and What People Actually Do. ESOMAR 3D Conference 2012, Amsterdam

0.62

0.56

Regularly

i.e. rank 2, 3…

Past 1 month

0.96 0.96 0.96 0.98

0.93

Aggregate-level correlation with spend