This document describes a new method from TNS for predicting consumer choice and optimal pricing strategies. The key points are: 1) TNS has developed a behavioral framework that puts buying decisions into context by considering consumers' habits, price knowledge, brand perceptions, and other individual context data. 2) This framework is used in an intelligent interviewing technique that designs choice scenarios tailored to each respondent based on their past responses. This provides better data from shorter, simpler interviews. 3) Market simulations run using the new behavioral choice model more accurately predict how consumers will respond to price changes compared to conventional choice models. The new method reduces deviation from actual market shares by 22%.

1. ©TNS
TNS Value Manager
A new method to predict
consumer choice for
optimal price strategies

2. ©TNS
Overall, Russian consumers are…
More money
conscious
Less happy
about
standards of
living
Are worried
about the price
increase
1

3. ©TNS
The shoppers are becoming more money conscious…
3
50%
58%
49%
43%
25%
35%
45%
55%
65%
2008/1 2009/1 2010/1 2011/1 2012/1 2013/1 2014/1 2015/1
% of population
+8pp
Silly to spend money
on luxury goods
–6pp Do not go to cheap stores
2015 vs. 2008
Source: TNS, Marketing Index / TGI, Russia (cities 100+), Base: 18+ y.o.

4. ©TNS
…and are better planning their purchases
4
48%
50%
37%
29%
47%
56%
25%
35%
45%
55%
65%
2008/1 2009/1 2010/1 2011/1 2012/1 2013/1 2014/1 2015/1
% of population
+2pp
Make no unexpected
expenditures
–9 pp
Spend money
without thinking
2015 vs. 2008
+9pp
Always read carefully
the product ingredients
Source: TNS, Marketing Index / TGI, Russia (cities 100+), Base: 18+ y.o.

5. ©TNS
Claiming changes in consumption
people tend to save money in different ways:
5
reduce budget… …or leave categories
Focus on promo
and sales
Change price
segment
Lapse some
categories
Switch to DIY

6. ©TNS
Building a pricing system around better data
6
"When we started to take pricing
seriously, we realized that we
needed not only to collect
better data but to be more
systematic about using it.“
(Petr Partsch – sales chief for Linde Gases,
Czech Republic)

7. ©TNS
Better data with Technology Enabled Research
7
We’re listening,
but are we learning?
We’re collecting data,
but are we listening?
/ ˈlɪsn / [verb]
to hear what someone has
said and understand that
it is serious and important
/ lɜːn / [verb]
to improve your behaviour as
a result of gaining greater
experience or knowledge of
something

8. ©TNS
A behavioural framework of consumer choice
8By 2017, 87%of connected devices will be smartphones or tablets
Behavioural
framework of
consumer choice
Choice
Model
Intelligent
interviewing
Behavioural
Economics
The expanded model puts buying
decisions into context:
 Consumers actively “build”
their choice sets
 Buying habits, price knowledge
and brand images inform sets
and choices

9. ©TNS
Intelligent communication with active consumers
9
...
Task 1
Interviewee
makes her
choice
New scenario
is designed
on the fly
Task 2 Task 3
utility
balance
utility
balance
Individual
context data
- habits
- knowledge
- perceptions
Interviewee
makes her
choice
Design algorithm
computes a tailored
choice scenario with
relevant products
New scenario
is designed
on the fly

10. ©TNS
Two-fold innovation
10
Behavioural Economics
 Habits, heuristics and
cognitive limitations
of decision makers
inform the design
 Context information
is treated as signal,
not noise
Intelligent interviewing
 Rather than survey
passive respondents,
we interact with
individual consumers
 We get better data
from shorter, simpler
interviews

11. Market simulations

12. ©TNS
A case study
12
Market simulations are the key output of our pricing tool:
How strongly will buyers react to new pricing scenarios?
In our case study we look at the German beer market in
the first 26 weeks of 2015.

13. ©TNS
Weekly market simulations for Krombacher 20x0.5 ltr
13
w1 w2 w3 w4 w5 w6 w7 w8 w9 w10 w11 w12 w13 w14 w15 w16 w17 w18 w19 w20 w21 w22 w23 w24 w25 w26
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
market share

14. ©TNS
Weekly market simulations for Krombacher 20x0.5 ltr
14
w1 w2 w3 w4 w5 w6 w7 w8 w9 w10 w11 w12 w13 w14 w15 w16 w17 w18 w19 w20 w21 w22 w23 w24 w25 w26
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
market share
conventional method
Conventional choice models tend to overestimate the impact of price changes

15. ©TNS
w1 w2 w3 w4 w5 w6 w7 w8 w9 w10 w11 w12 w13 w14 w15 w16 w17 w18 w19 w20 w21 w22 w23 w24 w25 w26
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
market share
conventional method
new method
Weekly market simulations for Krombacher 20x0.5 ltr
15
Mean
deviation is
reduced by
22%
The new method predicts the impact of price changes on
consumer choice much more accuracy

16. ©TNS
Listen to each respondent and learn
16
Our behavioural framework implies that the individual
context of choice matters:
 Does the individual respondent know prices?
 Which brands are in her purchase repertoire?

17. ©TNS
The role of price knowledge
17
We expect that buyers with little or no awareness of
prices in the category are less price-sensitive.
Do we find that in our market simulations?

18. ©TNS
The role of price knowledge
18
9.99 € 10.99 € 11.99 € 12.99 € 13.99 €
with awareness
without awareness
Conventional method
9.99 € 10.99 € 11.99 € 12.99 € 13.99 €
New individually adaptive method
25%
Differential impact leads to realistic predictions of price elasticity
Krombacher changes
in price, all others
remain constant
5%

19. ©TNS
The role of brand repertoires
19
We expect that consumers do not change their habits easily.
They will be price-sensitive if a particular brand is in their
purchase repertoire. If they never bought a brand in the
past, they will ignore price changes of that brand.
Do we find that in our market simulations?

20. ©TNS
The role of brand repertoires
20
9.99 € 10.99 € 11.99 € 12.99 € 13.99 €
occasionally bought
never bought
9.99 € 10.99 € 11.99 € 12.99 € 13.99 €
25%
Krombacher changes
in price, all others
remain constant
Differential impact leads to realistic predictions of switching behaviour
Conventional method New individually adaptive method
5%

21. ©TNS
Better data in a mobile world
21
By 2017, 87% of connected devices will
be smartphones or tablets
When interviewed on mobile devices
50%
interrupt the interview at least once.
After 10 minutes on a smartphone
40%
drop out.
Mobile devices need shorter, smarter
communication.

22. ©TNS
Fewer data in a mobile world
22
Smaller displays,
shorter attention spans
Fewer questions,
fewer data per respondent

23. ©TNS
Weekly simulations for Krombacher 20x0.5 ltr
23
w1 w2 w3 w4 w5 w6 w7 w8 w9 w10 w11 w12 w13 w14 w15 w16 w17 w18 w19 w20 w21 w22 w23 w24 w25 w26
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
market share
new method PC/Laptop
new method MOBILE
…Simulations from mobile data are slightly less accurate,
but still very close to real market shares

24. ©TNS
Building a pricing system around better data
24
Behavioural Economics
 teaches us that context
matters: Habits and
heuristics inform
buying decisions
 helped us to build a
behavioural framework
around our choice
models
Intelligent interviewing
 Saves time and gives
us better data
 enables mobile
surveys even for
demanding prediction
tools