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

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

The enormous power of winning 1 1 2 3 4 5 2 3 4 5 This presentation is about how these inequalities emerge

“ One is wonderful. Two is terrific. Three is threatened. Four is fatal ” ~ Larry Light

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

“ “Market share versus rank across 506 food ” brands and 665 sporting-goods brands Power laws! Source: Kohli, R and Sah , R. 2003. “Market Shares: Some Power Law Results and Observations”, Harris School Working Paper, Series 04.1

Man can be thought of as perfectly rational John Maynard Keynes

                       

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

Nope. Humans have bounded rationality Herbert Simon

Markets aren’t fair

We make good enough choices from a ‘stacked deck’ Sources: http://www.pepsico.com/Annual-Reports/2008/performance/n-america-csd.html

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

Usage of brand Frequency of purchase Newspaper 1 65.1 2.5 38.9 1.2 Newspaper 2 28.2 0.9 Newspaper 3 25.2 0.7 Newspaper 4 24.0 0.7 Newspaper 5 17.0 0.5 Newspaper 6 Source: UK newspaper market

            We do the best we can within our cognitive limits

      Get the important bits right and the rest will follow

The brain is impressive…

The brain is impressive… but it does have limits

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

Brand Brand Brand Brand Brand Brand Consider Brand Brand Brand Used Brand Used Brand Used Brand Brand Consider Brand Brand Brand Brand Brand Brand

88 1,058,290 countries respondents 205 1,267 categories studies

So, just how many brands do we realistically hold in mind at a time? Mean 3.9 Median 3.0 2.0 Mode Standard deviation = 3.7 Data: 1,267 studies, 88 countries, 205 categories, 1,058,290 respondents

% of observations 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 Evoked set size Standard deviation = 3.7 Data: 1,267 studies, 88 countries, 205 categories, 1,058,290 respondents

Consumer Technology Business Finance Energy Healthcare Auto Polling Services & Social 5.0 3.4 3.3 3.3 3.1 3.1 2.3 2.3 Standard deviation = 3.7 Data: 1,267 studies, 88 countries, 205 categories, 1,058,290 respondents

Beer UK Germany France Poland Spain Cambodia Thailand India 10.1 8.7 8.5 8.5 8.0 3.7 2.9 2.8

100 90 80 Linear R 2 Correlation 0.13 0.36 70 60 50 40 # brands in brand list 30 20 1,267 studies 10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Average evoked set size Standard deviation = 3.7 Data: 1,267 studies, 88 countries, 205 categories, 1,058,290 respondents

Evolution made us care about

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

Sources: http://xkcd.com/1098

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

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

Power law! R² = 0.97 Share of wallet (panel data) 69.5 23.6 11.3 4.1 1 2 3 4 Share of wallet rank Rank 1 brands get most of the share n=984 UK laundry detergent | Actual panel data supplied by KWP

Rank 1 Rank 2 Rank 3 % of observations Rank 4 90% 80% 70% 60% 50% 40% 30% 20% 10% 0 Share of wallet UK laundry detergent | Actual panel data supplied by KWP

log y y x log x

Power law! Power law! Caption: “Probability of visiting a merchant, as a function of merchant visit rank, aggregated across all individuals. Dashed line cor respond 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 Source: Krumme, C, et al (2013) The predictability of consumer visitation patterns

Power law! R² = 0.93 Share of wallet (panel data) 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 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 Performance rank Top rated (i.e. rank 1) brands get most of the share UK laundry detergent | Actual panel data supplied by KWP

It is possible to measure too much information

Respondent-level correlation with spend Awareness metric First mention 0.57 i.e. rank 1 All unaided aware 0.28 i.e. rank 2, 3… Aided aware 0.17 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

Respondent-level correlation with spend Usage metric Most often 0.71 i.e. rank 1 Regularly 0.62 i.e. rank 2, 3… Past 1 month 0.56 Past 3 months 0.58 Ever bought 0.37 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

# attributes in survey 78 36 36 36 Cell 1 Cell 2 Cell 3 Cell 4 Less measurement; richer data

# attributes selected 78.0 17.0 18.0 28.0 Cell 1 Cell 2 Cell 3 Cell 4

# ticks per respondent 2.2 3.0 3.2 4.8 Cell 1 Cell 2 Cell 3 Cell 4

Total survey time 13.0 3.0 3.5 5.5 Cell 1 Cell 2 Cell 3 Cell 4

Conclusions       Humans have bounded rationality

3.9

1 2 3 4

First Most Less is mention often more

Data Respondent Time Money

Thank you Acknowledgements: Anna Elanie de Constantin Jannie Ken Bell Bruno Retief Beer Michael Hofmeyr Gonçalves

Respondent-level Aggregate-level correlation with spend correlation with spend Awareness metric First mention 0.57 0.92 i.e. rank 1 All unaided aware 0.28 0.81 i.e. rank 2, 3… Aided aware 0.17 0.68 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

Respondent-level Aggregate-level correlation with spend correlation with spend Usage metric Most often 0.71 0.96 i.e. rank 1 0.96 Regularly 0.62 i.e. rank 2, 3… 0.96 Past 1 month 0.56 0.98 Past 3 months 0.58 0.93 Ever bought 0.37 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