Exploring
the ways tech savvy consumers have innovated in their lives and how market
research practices can follow their lead.
Innovation is
impacting every part of our lives, with smartphones making the world available
from consumer pockets with a simple click, swipe or voice command. And now
Apple HomePod and Amazon Echo, teamed up with Siri and Alexa, are providing
direct access to information without the need to reach for a device.
In fact,
voice-control products are taking off. According
to Kantar’s latest research findings, 21% of people now own a smart speaker or
home assistant. Given that the
first smart speaker (the Amazon Echo) was only launched in 2014, this indicates
that the voice-activated smart device revolution is just getting started. And
while these devices become ubiquitous, smartphones are still the central place
that consumers store everything, make mobile payments, shop online and more.
The momentary panic when you think you’ve lost your device is gut wrenching
because of the heavy dependence on them to function even in the most basic
sense. And it’s evolving at an exponential rate.
But can the same be said for
market research? Technological development in the market research industry is
often focused on how a business can create the latest programme or app to
benefit their business. However, the fact that technological development is
becoming a technological dependence for consumers means market research is
already benefiting. Or rather, it should be. The opportunity exists, given
access to data from new sources like customer data platforms, conversational
AI, and all corners of the internet but now researchers need to tap into this
wealth of data to better understand today’s consumers.
Recognising this, Kantar has
been exploring what the industry can learn from advances in and adoption of
consumer technology, and how to maximize insights and personalise experiences
through understanding connected consumers and the digital landscape they’re
building.
So, let’s explore some of
the opportunities within our industry.
Searching
for answers that are already there
We’ve all been there, a
conversation at a dinner with friends where we can’t remember the name of the
actress in that film or who won the finals last year. It’s tough to imagine
what we did in the days when we couldn’t simply “Google it” or ask Alexa for
the answer without pause.
Having any and every answer
at your fingertips is an efficient, useful tool. If an answer exists, then
isn’t it better to know and be able to move on to the rest of the conversation?
Think about research in the same way to look for what could already be out
there. Researchers are accustomed to creating a survey that covers all the
queries they have. However, there is an abundance of data already available
from others having asked the same questions. So why not search as an initial
step in the research process?
Not only does leveraging
collective survey knowledge provide instant answers, it can enable the
reduction of questions needed in subsequent surveys – saving research budget
and gaining meaningful insight from a more tailored survey that succeeds it.
Being able to tap into data
that already exists offers a significant advantage for the marketing industry.
This capability allows brands, marketers and researchers alike to leverage
audience understanding, quickly and cost-effectively by using real,
permission-based data without even asking a question. The beauty lies in the
simplicity and speed of searching what’s already been collected and then
connecting this into a larger research project or directly into strategic
action.
Connecting
multiple consumer datasets
Survey research is a
critically important method of tapping into the attitudes and opinions of
consumers, but respondents have a digital life outside their online panel
community – as a consumer, an influencer and a potential buyer.
Consumers are leaving a
bigger digital fingerprint that increases exponentially by the minute, if not
by the second. It expands by source, by type and size. It is creating an
ocean-sized pool of third-party data, which we are still learning how to best
utilise in data collection and analysis.
Consumers are leaving a bigger digital fingerprint that increases exponentially by the minute, if not by the second.
How can we connect to and
access this ocean of data? How do we link survey data to it for a holistic,
enriched view of our customers and non-customers?
Connecting data across
first, second and third-party sources is both an art and science. There’s even
the concept of zero-party data, a subset of first-party, which refers to
attributes a customer intentionally and proactively shares with a brand.
There are three main ways to
connect available data sources, privately and compliantly.
1. By person: The
first way to match an individual or their household to a data segment or index
is via Personally Identifiable Information (PII). Name, address, postcode,
mobile number, email address, are all data points that can be used to make the
connection. This is called a deterministic match.
2. By tagging: Second
in line is cookie drops or pixel tags. This match is done on a device level,
where a cookie is dropped on a computer, or a pixel tag is implemented in an
app on a smartphone to track behaviour and build a profiling data set around
who a consumer is. This is called a probabilistic match.
3. By device: Finally,
there is Device IDs. This is a match at a device level and driven by smartphone
technology. Every Android and iPhone has a Device ID that is created as part of
a Google Account or Apple ID. Originally conceived for delivery of targeted
digital advertising, it has become an essential driver in identifying profiles
between data sources. It’s still a probabilistic match like cookies and pixel
tags, but it reduces uncertainty.
The ability to connect
profiles through connection approaches is key to enhancing datasets. Once done,
it unlocks richer, more actionable data for targeting or even media activation.
Survey
segments for programmatic advertising
“Hey Siri, what’s the
address of the airport?” Although we can’t order an Uber to the airport through
a survey-based research project, we have seen how innovation in platform
connectivity and enrichment has increased the value of the research we are
delivering to clients each day.
In addition to the more
obvious ways that data and connectivity increase the richness of our research
and the speed with which we deliver, we now have more ways of making our
deliverables actionable – such as utilising the segments derived through survey
research with programmatic advertising.
In the past, the linkage
between customer segmentation and media buying or ad targeting was indirect at
best. Media planners and media buyers would use proxies such as demographic
profiles of their customer segments to indirectly target ads digitally (e.g.,
“my segment is more heavily weighted to females, age 26-34 in the Northeastern
US”).
As programmatic ad buying
has become the norm and more data has been introduced into the ad targeting
process, we can now more directly use custom segment data to display ads to
consumers who are more likely to fall into the segment (or audience) of
interest.
So, how does
it work?
Survey and segment: Conduct
research among respondents who have agreed to participate in this use-case and
who are connected into a Data Management Platform (DMP). Use the survey data to
define a segment of interest and identify the specific survey responders who
are in-segment (usually n=500 to n=1,000).
Scale: Using statistical
techniques and the data housed in the DMP, data scientists/modellers will
identify many more online profiles that are similar to the survey responders
identified via the survey (this larger group is known as a “scaled audience”).
Distribute (target and activate):
After ensuring that the survey responders are removed, the anonymous
identifiers of the scaled audience (e.g., cookie IDs or mobile ad IDs) are made
available to media buyers via the programmatic advertising process. Access to
the scaled audience is typically paid for by the advertiser based on the number
of ad impressions served.
Measure: Execute brand-lift
or sales-lift research to measure the results of the targeted campaign.
Brands are looking for every
possible way to make their digital marketing dollars go further, and custom
survey-based audience targeting offers this value add. We’ve seen incredible
results from custom research-based audience activation across industry
verticals, from consumer goods where we have measured a 400% increase in product
sales, to a 500% increase in site bookings for a travel industry client, to
double-digit increases in brand conversion across other FMCG and specialty
retail. By taking action on the wealth of data captured through research
projects, brands can now ensure that digital audiences take into account those
all-important consumer attitudes about a brand and its products.
Using AI to
create a conversation
Research innovation is
moving into AI. And the key for AI in market research is in conversation;
striking a conversation in an environment where respondents are already
familiar and active allows the research to blend seamlessly into the medium so
they feel like they are communicating with a friend. By doing this it opens the
door to more personal, open and honest responses. Through a dynamic script a
question can evolve to build on an idea in real-time and develop a deeper
understanding on a topic, allowing a richer view of audiences.
Chatbots are one way to
provide the opportunity for respondents to participate in a more interactive,
conversational manner than a typical typed response and in turn can deliver
deeper understanding and richer data.
For example, imagine
conducting a diary study using conversational AI. The respondent can be in the
kitchen testing the product as they make dinner and record an in-the moment
voice response – a convenient way for them to go about their day while
providing brands feedback. And it’s not only about convenience; a qualitative
and quantitative hybrid approach like this allows the researcher to obtain
greater detail. Not only does speaking a response tend to provide additional
information than typing, details can come in the form of images and videos from
people in the moment — a self-reporting data enrichment that is particularly
useful for diary studies seeking real-life behaviours.
And AI continues to learn
and improve, meaning the way we conduct research can too.
Connecting to the innovative
consumer and leveraging vast, new sources of data tied to them is key to
building richer, more useful audience profiles. As consumers and their tech
habits evolve, it’s imperative market research does too – this is something we
can learn from Siri and Alexa and from the audiences we are seeking to
understand.