This document provides an overview of a masterclass session on finding the story in data. It discusses the importance of using frameworks to structure data analysis and produce useful insights and stories. Specific frameworks elements discussed include defining the problem, establishing existing context, and assembling evidence from various sources. The document also covers what constitutes an insight, finding the main story and relevant exceptions in the data, techniques for data reduction, and using various visualizations and methods to craft narratives from the data. As an example, it analyzes survey mode usage data across several Asia-Pacific countries to identify trends and differences between countries.

1. Finding the Story in the Data
A masterclass session with
Ray Poynter
Founder of NewMR.org, UK

2. Agenda
• Frameworks
• What is an Insight
• Finding the Main Story
• The Relevant Exceptions
• Data Reduction
• Global Market Research example
• Q & A

3. The Data Do Not
Speak for Themselves

4. 3 Steps
1. Find the story
2. Craft the story
3. Communicate the story

5. Data to Insight to Wisdom
https://random-blather.com/2014/04/28/information-isnt-power/

6. Frameworks
A structured approach to finding the
insight in the data

7. Frameworks
Most of the teams that reliably produce good analysis
and useful stories use frameworks
– Individuals are less dependent on frameworks
Elements of frameworks
– How to frame the problem
– Linking the project to a wider context
– A standard method of organising the data (qual and
quant)
– Systematic methods of analysing data
– A preferred method for extracting the story and linking
it to the wider context

8. Define the Problem
“A problem defined is a problem half-solved”
Sources of information:
– The request for a study
– The proposal
– Discussions between sponsor, insight team and supplier
• What is the background to the project?
• What would success look like?
• What actions should follow from the research?
• What do people think the results are going to be?
(Or, what are the prevalent hypotheses?)
Smith & Fletcher, 2004

9. Establish What is Already Known
• The frameworks approach avoids focusing on
just the current research project
• The analysis, the validity, and the story need
to blend research with the wider context
• The context is a web of existing knowledge:
– Within your organisation
– Within the agency/supplier
– In the public realm

10. What we need to know?
Known
‘Knowns’
Known ‘Unknowns’
Unknown
‘UnKnowns’
Check these are valid
Plan how to gather these
Expect the unexpected “It ain't so much the
things we don't know
that get us in trouble. It's
the things we know that
ain't so.”
Artemus Ward

11. Many Insights Arise from the Unexpected
Look for the Unexpected

12. Who is the project for? _________________
What is the business issue/problem that is being addressed?
__________________________________________________
What does the business want to do, once it has addressed this issue?
______________________________________________________
What do we already know?
Item Held by: Description
1 ______ ______ ______________
2 ______ ______ ______________
3 ______ ______ ______________
Assumptions and predictions
Who What
1. ______ ______
2. ______ ______
Simplified

13. Assembling the Evidence
• Quantitative
– Standardize? Missing Data? Indexing?
Re‑basing?
• Qualitative
– Translations? Transcripts? Notes?
• The nature of the sources
– Credibility? Bias? Interactions?

14. What Is An Insight?
• New
• Different way of looking at things
• Facilitates new outcomes
• Not tied to a specific situation - reusable
• Often obvious afterwards
• Useful

15. 5 Cs of Insight
Connections
Coincidences
Curiosities
Contradictions
Creative jump
} Build on the
current picture
Rethink an assumption
Discard an assumption
HT, John Storey

16. Use the Business
Question as a Lens
• We are not trying to find every story in the
data
• We are not trying to find the most
interesting story in the data
• We are finding a story that helps solve the
business question

17. Finding the Story
• Know what the question is.
• Have an idea of what success looks like.
• What is the main story?
– What do most people do? Why do most people do it?
• What are the relevant exceptions?
• Determine how the message in the data answers
the business question and craft that as a story.

18. The Lead
Nora Ephron
When Harry Met Sally
Sleepless in Seattle
1st Day in Journalism School
5 Ws (Who, What, When, Where & Why?)
Asked to write the Lead for the school newspaper
“The entire school faculty will travel to Sacramento next
Thursday for a colloquium in new teaching methods.
Among the speakers will be anthropologist Margaret
Mead, college president Dr. Robert Maynard Hutchins, and
California Governor Edmund Brown.”
All the students wrote about the 5Ws – good, but not
right.
The Lead?
No school next Thursday!

19. Data Reduction
Remove anything that is not David

20. Making Numbers More Visual
• 2 or 3 SD – significant digits
• Sorting
• Indices (Indexes)
• Colour Coding

21. Per Capita Govt. Spend on Health - $
183 Countries, 1995 to 2010, source Gapminder

22. 3SD – Units of $10
Divide by 10, display no decimals (but don’t delete them)

23. Sorting – Filters in Excel
Adding an ID field let’s you get back to the starting position easily

24. Looking at the Shape of the Data

25. Index, 1995 = 100

26. Line Chart for Outliers
Congo,
Dem Rep
Niue
Georgia
Equatorial
Guinea
Sudan
$ by Govt on Health
Indexed 100=1995

27. Using ‘Cut Up’ for Narratives
Dadaism
William S Burroughs
Beat poet & artist David Bowie

28. Using ‘Cut Up’ for Narratives
Mind mapping software can also be used
For each data table, chart, or
PowerPoint page – find the main
story.
Write this story on a card – perhaps
with colour coding.
Move the cards around to find the
story. Similar cards close together.
Missing links suggest more analysis
required.
Build the story architecture.

29. Example
1. Can I use online surveys in all
of the main APAC markets?
2. Where are online surveys
strongest?
3. How different is the picture
from one country to
another?

30. Framing and Scoping
Can I use online surveys in all of the main APAC markets?
• What do we mean by APAC?
• What do we mean by Surveys?
• What do we mean by Online?
• What do we already know?
• Predictions – Which countries most online?

31. What do we mean by APAC?
Turnover US $M Turnover US $10M
Japan 1891 189
China/ Mainland 1884 188
Australia 791 79
South Korea 464 46
India 269 27
Singapore 173 17
Thailand 172 17
China/ Hong Kong 141 14
Indonesia 122 12
Malaysia 108 11
ESOMAR GMR 2017

32. What do we mean by APAC?
Turnover US $M Turnover US $10M
Japan 1891 189
China/ Mainland 1884 188
Australia 791 79
South Korea 464 46
India 269 27
Singapore 173 17
Thailand 172 17
China/ Hong Kong 141 14
Indonesia 122 12
Malaysia 108 11
ESOMAR GMR 2017

33. Predictions
• Which of these countries will be the
highest, in terms of share of surveys online?
• Which countries will have the lowest
proportion of their surveys online?
• What will be the highest mode that is not
online?

34. What do we already know?
Country
Internet
Penetration
GDP PPP per
capita Int$1000
Japan 93% 43
South Korea 93% 39
Australia 88% 50
Hong Kong 87% 61
Singapore 84% 91
Thailand 82% 18
Malaysia 78% 29
China 55% 17
Indonesia 54% 12
India 34% 7
Internet: InternetWorldStats, GDP: IMF, r2 between these = 43%

35. Link Between Internet & Income
0
10
20
30
40
50
60
70
80
90
100
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
GDP per capita Int $1000
Internet Penetration
GDP per capita versus Internet Penetration
Once a country
reaches about
$17K per person,
the penetration
is over 75%
The r2 between internet penetration and the log of per capita income is 72%

36. What do we already know?
Online is the cheapest mode.
Online is faster than CATI, F2F &
Mail.
Reasonable to assume online is
preferred when suitable

37. What do we mean by
Online & Surveys?
Country %
Quantitative
research
Qualitative
research Other Total
Australia 80 15 5 100
China 75 24 1 100
Country % Mail CATI F2F Online Mobile
Online
traffic
Auto-
mated Other
Total
quant
Australia 0 6 2 44 12 12 4 0 80
China 1 11 29 13 0 0 0 21 75
ESOMAR GMR 2017

38. What do we mean by
Online & Surveys?
Country % Mail CATI F2F Online Mobile
Online
traffic
Auto-
mated Other
Total
quant
Australia 0 6 2 44 12 12 4 0 80
China 1 11 29 13 0 0 0 21 75
Country % Mail CATI F2F Online Mobile
Total
Survey
Australia 0 6 2 44 12 64
China 1 11 29 13 0 54
Country % Mail CATI F2F Online Mobile
Total
Survey
Australia 0 9 3 69 19 100
China 2 20 54 24 0 100

39. Surveys
Country % Mail CATI F2F Online Mobile Total
Australia 0 6 2 44 12 64
China 1 11 29 13 0 54
Hong Kong 0 22 14 28 6 70
India
Indonesia 0 6 60 3 8 77
Japan 7 1 7 50 1 66
Malaysia 1 20 25 20 1 67
New Zealand 2 11 8 44 0 65
Singapore 0 15 8 35 5 63
South Korea 1 18 23 24 4 70
Thailand
ESOMAR GMR 2017

40. Surveys
Country % Mail CATI F2F Online Mobile Total
Australia 0 9 3 69 19 100
China 2 20 54 24 0 100
Hong Kong 0 31 20 40 9 100
Indonesia 0 8 78 4 10 100
Japan 11 2 11 76 2 100
Malaysia 1 30 37 30 1 100
Singapore 0 24 13 56 8 100
South Korea 1 26 33 34 6 100
ESOMAR GMR 2017

41. Surveys
Country % Mail CATI F2F Online Mobile
Australia 0 9 3 69 19
China 2 20 54 24 0
Hong Kong 0 31 20 40 9
Indonesia 0 8 78 4 10
Japan 11 2 11 76 2
Malaysia 1 30 37 30 1
Singapore 0 24 13 56 8
South Korea 1 26 33 34 6

42. Surveys
Country % Mail CATI F2F Online Mobile
Australia 0 9 3 69 19
China 2 20 54 24 0
Hong Kong 0 31 20 40 9
Indonesia 0 8 78 4 10
Japan 11 2 11 76 2
Malaysia 1 30 37 30 1
Singapore 0 24 13 56 8
South Korea 1 26 33 34 6
Conditional Formatting

43. Sort the Columns
Country % Online F2F CATI Mobile Mail
Australia 69 3 9 19 0
China 24 54 20 0 2
Hong Kong 40 20 31 9 0
Indonesia 4 78 8 10 0
Japan 76 11 2 2 11
Malaysia 30 37 30 1 1
Singapore 56 13 24 8 0
South Korea 34 33 26 6 1
Online is the most prevalent
F2F is the clear 2nd largest overall

44. Sort the Rows
Country % Online F2F CATI Mobile Mail
Japan 76 11 2 2 11
Australia 69 3 9 19 0
Singapore 56 13 24 8 0
Hong Kong 40 20 31 9 0
South Korea 34 33 26 6 1
Malaysia 30 37 30 1 1
China 24 54 20 0 2
Indonesia 4 78 8 10 0
Online largest in 4 countries
F2F largest in 2 countries
Online and F2F similar in 2 countries

45. Main Online Story
Country % Online F2F CATI Mobile Mail
Japan 76 11 2 2 11
Australia 69 3 9 19 0
Singapore 56 13 24 8 0
Hong Kong 40 20 31 9 0
South Korea 34 33 26 6 1
Malaysia 30 37 30 1 1
China 24 54 20 0 2
Indonesia 4 78 8 10 0
Japan= No.1
Australia, 2nd
Indonesia almost 0%

46. Subplots
Country % Online F2F CATI Mobile Mail
Japan 76 11 2 2 11
Australia 69 3 9 19 0
New Zealand 68 12 17 0 3
Singapore 56 13 24 8 0
Hong Kong 40 20 31 9 0
South Korea 34 33 26 6 1
Malaysia 30 37 30 1 1
China 24 54 20 0 2
Indonesia 4 78 8 10 0
F2F: Indonesia (high) and Australia (low)
CATI: Japan (low)
Mobile: Australia & Indonesia – 2nd largest mode
Mail: Japan relatively high

47. Typical APAC Countries?
Difference Online F2F CATI Mobile Mail Total
Japan 31 18 17 5 9 80
Australia 24 26 9 13 2 74
Singapore 11 16 5 2 2 37
Hong Kong 4 9 13 3 2 31
South Korea 10 4 7 0 1 22
Malaysia 15 8 11 5 1 39
China 20 25 2 6 0 53
Indonesia 41 49 11 4 2 107
Mean 44 29 19 6 2
Difference = Absolute value of mean – score

48. Typical APAC Countries?
South Korea, HK & Malaysia = typical for APAC
Australia/Japan online
Indonesia F2F
Difference Total Online F2F
South Korea 18 34 33
Hong Kong 29 40 20
Malaysia 35 30 37
Singapore 41 56 13
China 49 24 54
Australia 78 69 3
Japan 86 76 11
Indonesia 101 4 78

49. What about Thailand?
Can I use other data to give me an estimate?
Country
MR
turnover
in 2016
(US$m)
Adspend
in 2016
(US$m)
Population
in 2016
(m)
MR spend
per capita
(US$)
Ad spend
per capita
(US$)
MR as
% of ad
spend
Thailand 172 3246 69 2 47 5%

50. Online Surveys Model
Country
(A) Internet
Penetration
(B) GDP per
capita
(C) MR
Turnover % Online
Japan 93 43 189 76
South Korea 93 39 46 34
Australia 88 50 79 69
Hong Kong 87 61 14 40
Singapore 84 91 17 56
Malaysia 78 29 11 30
China 55 17 188 24
Indonesia 54 12 12 4
Lin Regression
Online% = 0.81*A +0.4*B +0.16*C -50.9
R=0.93, R2=87%

51. Thailand?
Country
Internet
Penetration
GDP per
capita
MR
Turnover Estimate % Online
Thailand 82 18 17 26 NA
Taiwan 88 50 9 42 34
The estimate for India is close to 0% – which is where
it should be for gen. pop. or nat. rep. – IMHO

52. The Story
Review The Question!

53. Example
1. Can I use online surveys in all
of the main APAC markets?
2. Where are online surveys
strongest?
3. How different is the picture
from one country to
another?

54. The Story
1. Online isn’t dominant everywhere in APAC for
mainstream projects
– In a ‘typical’ APAC country F2F and online are still
competing
– Japan & Australia very online, Indonesia very F2F
– Consider target group when choosing mode, consider
mixed mode
2. Mobile not following the same pattern as Online
– Australia AND Indonesia relatively strong

55. Thank You
Questions?
To find out more about Ray Poynter and NewMR visit NewMR.org