Marketing research-

an introduction to an
industry
IMT Ghaziabad
August 2024
The Taste of Tears - 1989

o Women from villages in Haryana were

UNICEF wished to know if Indian women could make invited to make a solution:
a safe & effective oral rehydration solution at home . 1. “According to your taste”
According to Lancet, oral rehydration therapy was 2. Matches with the “taste of tears”
“potentially the most important medical discovery of
the 20th century”. Salt plays a crucial role in o We found that none of the strategies
absorption, but the concentration needs to be correct: worked…
“India has achieved 86%
too little will not facilitate osmosis, too much salt can reduction in the number of
be dangerous. under-five diarrheal deaths from
1980 to 2015..”-Choudhary et al,
Journal of Global Health, 2019

- Hema Viswanathan and Jon E Rohde; formerly with IMRB and UNICEF respectively 2
Stages in research - for a “typical” study

Decide on an agency and

Defining the problem; the Prepare a brief: invite one or
research design…e.g.
decision(s) that we need more agencies to respond
quantitative, qualitative,
to take secondary etc.; sample
size

Analysis and data Data collection- via interview, Prepare questionnaire,

interpretation; report on focus groups, other means discussion guide, other
findings
materials

Action based on
recommendations/insights
Marketer
Researcher 3
Qualitative & Quantitative Research

What makes
Which of three
a pack pack designs
Qualitative appear Quantitative best convey
“premium”? “premiumness” ?
 Unstructured, flexible  Larger sample
 In depth  More structured
 No numbers (but….)  Numerical, quantification possible
 Can get behind barriers  Estimation/prediction possible
 Authenticity  Comparison within- sub-groups; time
Can be exploratory (first
stage), or complementary,
or standalone

EXPLORE UNDERSTAND EXPLAIN MEASURE ASSESS VALIDATE

© 2022 Borderless Access

 The origins of qualitative research- Ernest Dichter and “motivational research”

“Why do you use this “it’s not just the smell or Figure out the personality of
soap?:”.. one example price, but the gestalt of the product and you will
of a useless question” personality of the soap” understand how to market it

Smoking is symbolic
Baking is a symbol of
of virility…also a feminity…in a sense
legitimate excuse to like giving birth
..“” interrupt the day
..

Any comments on these statements?

Especially about “soap” 5
Types of research problems

Framing a
problem

6
Types of issues or problems
Monitoring, understanding, “strategy”

o Category size, growth, shares, distribution

o Audience measurement- TV, press, digital, other
o Brand health- e.g. continuous or other tracking; social media monitoring
o Customer satisfaction
o Segmentation and allied U&A (usage and attitude studies) or KAP (knowledge, attitude,
practices)

To monitor, to understand better 7

Types of issues or problems
To address a specific issue or query- some examples

o Should we launch this new product?

o Should we go with this ad (campaign)? Which of these two treatments will work better?
o Is this price point right- is my pricing strategy right?
o Does this pack convey what it should? Will it be noticed?
o Is there anything wrong with my product formulation? Can I replace this with something else?
What new flavours are acceptable?
o Our sales have plateaued. What should we do?

8
Some new trends in market research/ the use of technology in market research

9
How the TV metering system works in India

Audio “watermarks” in Received by “BARC” Viewership data

content that’s being telecast, households via a device transmitted to central
not audible to the human ear (microprocessor+modem) servers
(has Channel ID and time
attached to the TV set
stamp)

600+ watermarked channels Current sample = 55,000

households +1000+eateries

10
How can you get consumer response to ads, concepts, packs without
asking questions ?

Facial
recognition

Eye tracking

Neuroscienc
e

14
© 2022 Borderless Access
Eye tracking- a brief history

Louise Emile Javal reported in 1879: eyes don’t

move continuously along a written line of text but
they make short rapid movements, called
saccades, alternated to short stops, called
fixation.
Moreover, human eyes don’t stop on every single
word, but on some of them.
Readers don’t look at each word always, take the
missing information from the context…
How can eye-tracking be useful to a marketer?

o In a showroom- what interests the shopper

o What portion of the page in a website is
working
o Vehicles- perceived quality- where does the
eye fall
o What consumer will next look for- shelf or store
design or e-commerce maybe

16
Modern eye tracking
Can work with alternative devices

o Screen based- a special camera observes the o Best when we want to study response to
eye and software works out where the gaze something on a screen
might be
o Wearable- eye tracking glasses; with built in
o For the real world e.g. supermarket, consuming
screen camera and microphone
media at home, etc
o Wearable- virtual reality headsets
o Screen based- using the computer (or
mobile’s) own webcam- which is used to
calculate the position of the eye o When lower accuracy is okay and larger scale
testing is required

17
A simple example

Source: https://neilpatel.com/blog/eye-tracking-studies/

18
Facial coding- an introduction

o Paul Eckman (1960s) demonstrated that

Darwin was right- facial expressions
universally shared by humans, not dependent
on culture
o Examples of basic emotions conveyed by facial
expressions : happiness, surprise, fear,
disgust, sadness and neutral. Some theorists
add anxiety, confusion, anger etc. to the list
o Facial expressions are difficult to suppress,
they are spontaneous
The facial muscle that
o FACS* begins in 1970s with the definition of contracts when we smile-
facial movements- called AUs (Action Units) zygomaticus major (AU12)

*Facial Action Coding System

Coding facial expressions

o Facial expressions and emotions can be coded

by trained coders- or computer algorithms
o In automated facial coding, the first job is to
determine where the face is, then determine
the position of the main features (e.g. nose,
mouth, eyebrows, etc.), going on to track even
tiny movements in facial expressions
o Can work on laptop, mobile, etc

Why would a marketer or market

researcher be interested in facial
expressions?

20
Coding facial expressions- some examples

She seatselvisitors at aindesk,

her facing
small aoffice
monitor
o An AU consists of one (or sometimes more) Meeting Kaliouby …,
and a webcam; the software
Massachusetts, I see her contract her locates the
facial action visitor’s face and draws a box around it on the
zygomaticus major muscle, raising the
screen. It identifies a set of points
corners of her mouth, and her orbicularis to track: the
o An Emotion can consist of one or more than corners of the eyes and mouth,
oculi, crinkling the outer cornerstheof tip
her of the
eyes.
one AU. Some emotions consist of several AUs nose,
She isand so on. and
smiling, Twenty times each
I deduce that second,
she is
the software looks for “action
welcoming me, before she even gets outunits,” the often-
the
fleeting play
word “hello.”.. of muscles across the face. There
are 46 of these, according to the standard
Insincere smile Zygomaticus major
system of classification, the Facial Action
Coding System (FACS). They include inner and
Genuine smile Zygomaticus major, orbularis oculi outer brow raisers, dimplers, blinks, winks and
lip puckers, funnelers…. Affectiva’s standard
program samples about 15 of these at any
time, and analyzes them for expressions of
seven basic emotions: happiness, sadness,
https://imotions.com/blog/facial-action-coding- surprise, fear, anger, disgust and contempt,
system/#main-action-units plus interest and confusion
Source:https://www.smithsonianmag.com/innovation/rana-el-
kaliouby-ingenuity-awards-technology-180957204 / 21
Understanding Product Consumption- Smart Coasters

Combines IoT* devices with advanced analytics

These devices take the form of 'smart coasters' that
capture what, when and where a product is being
used, - & how much of it is used, in any room in the
home, in real time.
Aims to capture every usage occasion- where it
happens, when it happens, down to the product
type/SKU, gram and second

* Internet of things
Types of research problems

Research design- a
digression- on
questionnaire design*

* But crucially linked to problem statement 23

If you want a questionnaire that works…
We’ll t
ry
angles these two d
,t iff
o Seek just the right amount of getting wo different erent
at the w ays o
same f
information thing

o That enables you to take

decisions/action
We
al w
ke e p it, w e
as k ay s …let ’ s
e t ime to
h a v
que this don’t
stio argue
n

s w er al l
ey c a n an n s wer
If t h n a
e, t h ey ca
thes qu es tion…
er
ano t h
Key tests of a question:

Will the respondent understand the question? ‘Validity’ Are we measuring what we
think we are measuring?

Will the respondent be able to answer the

question?

Will the respondent be willing to answer the

question?

Will the answer really help you make ‘Actionable’’

decisions?* (Will we learn anything that we do not
know already?)

* Relevant for critical questions. Some questions may be there for other purposes
Types of research problems

Primary, secondary and

more

26
Primary versus Secondary research

Primary, “bespoke”- Conducted Secondary, “off the shelf”- what’s

afresh, customized for specific readily available: e.g. news ,
objectives government publications, syndicated
studies, etc. Usually there’s a cost-
sometimes free

A study to understand why some Determining viewership of various

TV serials do better TV serials

27
Some specialized, syndicated data sources

34
 Retail Audit
Retail Measurement Service (Nielsen)
 Some key metrics : Distribution

o Numeric Distribution : Number of outlets you are present in divided by total

number of outlets (that the category that category is present in)
o Weighted Distribution : measures the quality of distribution – if weighted
distribution is 80% it means 80% of the category sales happens in the outlets
you’re present
o Share amongst Handlers (SAH) : your market share in the outlets that you
are available in. For strong brands, this should be much higher than overall
share. Else you need to worry about other things -rather than availability.
E.g. If market share is 20% then SAH should ideally be around 28-30% -
unless ND is very high (say 85%+)
o Stock Turnover Days : depth of stocking of each SKU. Measured by dividing
closing stocks of an SKU by its average daily sales. This tells you how many
days’ worth of stock of that SKU is in the outlet Stock turnover should neither
be too high or too low.

36
Not all chains agree. Those
who do Coop and who do not
Non Coop. (POS = Point of
Sale)
Consumer Panel
Kantar Worldpanel
 Key features of the panel

o Sample of households recruited – agree to Over and above what you can get from the Retail
keep purchase records of the 100 odd Audit..(for example)
FMCG categories via purchase diary o Household penetration

o Diary data backed up by monthly interview; o Cumulative trial

collection of wrappers; pantry checks o Source of gain and destination of loss
o Loyalty
o Common sample every month
o Which brands bought along with my brand
o Brand share by SEC, etc

o Monitoring a large number of categories with a common sample provides a rich database
o To boost the ecommerce purchase pickup in Panel, have added ecommerce buying households as
booster households. Dual interviewing in these households to capture the ecommerce purchases from
the person who is actually doing it and rest purchases from the housewife

39
Click icon to add picture

Thank you

40