By Liesa - January 21, 2020
Whether you work in PR, marketing or consumer insights, the
scale and speed of available data has become so overwhelming that,
paradoxically, light-bulb moments are now few and far between. Predicting
upcoming trends, finding that magic content topic or acting on consumer opinion
too often turns into a frustrating, long-winded hunt for insights - success not
guaranteed. This is where Conversation Clusters come in.
What are conversation clusters?
Conversation Clusters is a user-friendly data visualization tool that helps you
instantly uncover, understand and visualize the context around any topic at a
glance. It enables you to find meaningful trends, patterns, content and
connections for any topic, no matter the volume of results.
At the same time, it eliminates all those hours spent on manual
research. By analyzing millions of results from media, consumer and social data
in under a minute, it maps hidden patterns within any conversation or topic so
you can focus on what’s important.
The best thing? There’s no need to be a data scientist to
generate or analyze Conversation Clusters - you’ll be extracting crucial
insights in a matter of minutes. Using the latest tech in machine learning
research, you can map out the conversations behind your search in a few clicks,
showing the context and connections behind your data points.
Let’s look at some examples to show you the potential of
conversational intelligence for various industries (automotive, FMCG, beauty,
telco, etc.).
1. Why macro trends are so important and how to leverage them
Identifying macro trends can help you understand context. Let’s
see how a music video trend can be an opportunity for the automotive industry.
No matter your line of business, you probably think you have a
pretty good grasp on what’s going on in relation to your brand and competitors.
You’ve set up alerts for your most important keywords, and you’re watching
social like a hawk.
But sometimes, even in your own industry, you might not be aware
of macro trends, and how you can use them successfully in your strategy. Brands
are increasingly looking to cash in on these trends, and there’s a simple
reason for it: talking about something that interests a lot of people and
sparks meaningful interactions is more likely to make a strategy or campaign a
success.
When you spend every day in the thick of it, seeing the forest
for the trees can become a problem. How can you identify these trends? After
all, millions of results and data points are only meaningful when you know
what’s going on and you can act on these macro trends. That’s when looking at
Conversation Clusters will work wonders.
Let’s look at a high-level example of Conversation Clusters for
the automotive industry.
At first glance, this network shows significant conversations
related to performance and car specs. There’s also a fairly large cluster about
owned brand marketing, as well as discussions about motor sports and
sponsorship.
Moving south of the center, there’s a high-volume cluster
containing articles about car accidents and crime, but which remains
disconnected from the rest of the conversations linked to this industry. In
between, there’s some discussion activity about corporate topics, recalls and
faults, and electric cars.
So far, so predictable.
Now, if you’re interested in any of these topics, you could
start a deeper analysis by zooming into specific clusters to get a more
granular understanding of the subject and its conversation drivers. But what
stands out at this level already, is the music industry cluster.
It takes a central spot within the conversation about the
automotive industry, suggesting it is connected to a lot of other topics around
it. Why are there so many conversations linked to music? Digging deeper reveals
that there’s a fairly large conversation bubble about cars that appear in music
videos. These conversations are garnering a lot of comments, and there’s a
number of cars that are regularly included in the lyrics of top hits.
How do you leverage product placements in music videos? How do
you target the most relevant audiences? What are other ways to relate priority
targets? Gather the insights that help inform your communication strategy, from
social media posts to more engaging advertising and different forms of content.
Once you have identified this macro trend, dig deeper with a
focused query to get more insights into other emerging trends in the industry.
2. How to predict trends for content creation
Predicting rising trends helps you find content ideas that
resonate. Check out why ugly fruits are a big thing in food waste
conversations.
Today, consumers care more and more about the origin of the
products they’re offered, as well as the far-reaching consequences of their
consumption choices. This is probably why the term “food waste” gained in
popularity on Twitter and beyond throughout 2019, with a total of 694,000
mentions over that time frame.
But knowing the wider topic alone isn’t enough to create
successful content. If you want your content to be timely, relevant and
original, you need to get into the business of trend prediction.
Let’s have a look at a map of this topic:
Several of the emerging themes will confirm your intuition when
it comes to food waste conversations, such as the interconnected topics of meal
planning and recipes for leftovers. Yet didn’t you expect these food prep
conversations to make up a bigger part of this cluster?
At the bottom, we find topics such as the confusion expressed by
consumers when it comes to the expiration date of their products, or questions
linked to the consumption of “ugly food” (imperfect food products such as
misshapen vegetables or bruised fruits).
If you intend to write engaging content on this topic, this data
visualization will help you understand the most engaging story angles, as well
as the debate drivers and the relationship between the different subtopics.
Leveraging data insights in your content creation will thus allow you to own
the conversations in relevant connected clusters which have not been spotted by
your competitors and help you establish your leadership in the topics you
choose.
3. How to use conversation trends and consumer insights to
improve your communication
Knowing what customers really care about elevates your
communications. Here’s how a major telco brand took its product launch to the
next level with the help of conversational intelligence.
You just got the go from your R&D department: your product
launch date is confirmed, and you still have a few weeks to tweak your
communication plan. How can you use this time to take your product launch from
good to great?
For a telco brand, the answer was using Conversation Clusters to
elevate their market research. Listening to what consumers are saying
about similar products, figuring out what their pain points are and paying
attention to shifting trends in opinions can be crucial in a successful launch
campaign.
In this case, in preparation for the launch of its new Internet
box, the brand’s customer insights team analyzed all market conversations
related to the product amongst its own customers, along with all direct and
indirect competitor boxes. What were the specific issues encountered by the
average consumer in their daily use of this type of device? How could these be
addressed throughout the product launch?
Filtering results such as showing only negative sentiment
discussions allowed the brand to zero in on specific pain points, while
analyzing positive clusters helped them refine their communication by
emphasizing features that people really loved.
Mapping the entire context of the discussion with Conversation
Clusters this way allowed them to improve the product sales briefs for the
brands’ vendors and the advertising materials, as well as to create FAQs and
other resources that addressed the most common issues that had been identified.
4. What trend research can do for more innovative product
development
If a trend analysis doesn’t take weeks, you can go to market
fast with a product that customers love. Such as this beauty brand that used
market insights to develop a unique product.
In the social media era, beauty fans are more informed than ever, and are
always on the lookout for their next favorite product. These consumers follow
influencers, discuss issues on forums, and share their comments and solutions
with their communities.
A globally recognized beauty brand used this wealth of
information to analyze business profiles from Instagram influencers from
different regions of the world, in order to forecast trends in foundation,
mascara, lipsticks and skin care products.
Using Conversation Clusters, they realized that the volume of
conversations about moisturizing products with added SPF on the European market
was lower than in other regions, and that customers had a specific problem with
the available products: stickiness.
Armed with that information, they developed a product with a
different and lighter texture that fit the insights they had extracted from
these conversations. Then they identified the local micro-influencers who had
been most vocal about the problem before and developed a campaign that included
sending them samples.
How do Conversation Clusters work?
Legacy query solutions take time to build and you need to know
what you’re looking for. The algorithm behind our Conversation Clusters can
group vast numbers of articles in minutes and find hidden connections that
might not be apparent to analysts from the outset.
“The algorithm identifies relevant word patterns and semantic
structures in the results and then plots them in a way that related
conversations are shown close to each other”, explains Dr. Benedikt Wilbertz,
Head of Data Science at Talkwalker.
“This way, it builds groups of discussions covering similar
subjects by discovering their common themes and showing how they relate to each
other. Using the latest advances in machine learning research and a massively
parallel infrastructure, we are able to analyze and represent tens of thousands
of conversations in little more than one minute.”
Here are a few tips which will allow you to extract even more
value from this data visualization:
·
Similar articles
cluster together to form bigger conversation dots.
·
Wide clusters are more
controversial or discuss broader topics.
·
Clusters in the middle
of the visualization are better connected with the overall conversation.
·
Niche conversations
happen on the edge of the map.
What is the difference between word clouds and Conversation
Clusters?
A word cloud identifies the most commonly used words or
expressions linked to a specific query. You may have some indications as to the
frequency at which these words are used based on the size of these words.
Conversation Clusters take your analysis to the next level by
mapping out the hidden connections between the different topics present in your
search. Rather than being summed up in a single word, the information is put in
context, and you have the option to zoom into the cluster to identify specific
articles.
Whether you want to find trends, write trending content, or use
consumer insights to make the best products, Conversation Clusters will help you predict future
developments that matter for your niche.
Liesa is part of the content team at Talkwalker (the loudest
part, some might say). Her mission is to help marketers navigate the listening
and analytics parts of their jobs. When
she's not doing research or writing, she can be found spending too much money
in the local bookshop. Follow her on Twitter or LinkedIn for everything
analytics and updates about her mustard obsession.
No comments:
Post a Comment