See the world in terrifying detail. Welcome to the world of BuzzWatch.

BuzzWatch, a buzz analysis tool that visualizes every detail of online buzz, is packed with features designed to solve the challenges of traditional social listening (such as complex and cumbersome operation).

Moreover, it goes beyond mere buzz analysis. It holds the potential to democratize and share the voices of individuals and specific points—often overlooked by companies and society—by highlighting: "People making these claims are gaining momentum!"

What exactly is the world of BuzzWatch, where you can see the world with frightening clarity? We delve into its true nature.

Are you accurately grasping that buzz?

The "buzz" happening daily on the internet has become incredibly important for people and companies living in the present. However, accurately grasping the reality of this buzz is quite difficult and requires significant time and effort. In short, it's too much hassle.

One method for analyzing SNS, the source of buzz, is utilizing social listening tools. We test various tools from around the world daily and appreciate their convenience. Yet, we also feel that traditional social listening tools alone offer "quite limited analytical perspectives" and "fall short of addressing the real pain points."

■Challenges of Conventional Social Listening Tools

(1) Many "missing perspectives" in data analysis
While they can analyze metrics like tweet volume, they often struggle to break down those mentions by "gender/age demographics," "positive/negative sentiment," or "emotional tone." They also seem weak at revealing the core content of the buzz itself, typically relying solely on a "keyword cloud" approach.

This leaves critical perspectives needed for "checking ad campaign reactions" missing. Understanding "who was talking about it, how, and how it spread" could only be estimated based on the analyst's gut feeling.

(2) Securing Resources That "Only Specialists Can Master"
Social listening tools become more complex to operate as their functionality increases. They also demand search and analysis skills, making them inherently labor-intensive and reliant on specific individuals. Currently, there is no tool that allows the average office worker to easily and accurately grasp buzz.

(3) Focused on "searching for individual terms"
Most traditional social listening tools focused on searching for individual terms. They required manually searching for specific company names, product names, individual names, or articles one by one.

While many companies and managers desire this, it often overlooks the crucial perspective of "the overall standing of the company or brand." Understanding whether buzz about one's own company or brand is high or low requires surveying the entire social landscape, not just competitors. Yet, users were often overwhelmed by localized searches, making it extremely difficult to gain this bird's-eye view.

To solve this challenge, we began research and development in 2015. To tackle the highly complex algorithm development, we collaborated with a Mongolian AI team (including Data Artist Inc.) that includes numerous Math Olympiad medalists, refining the system through repeated iterations. This culminated in the completion of "BuzzWatch," equipped with features that address those hard-to-reach pain points.

Now, let's take a look inside BuzzWatch—fun and super convenient!

BuzzWatch's Features: Stripping Buzz Bare

■CM/TV Program/Trending Words/Trending Tweets Rankings

BuzzWatch automatically delivers various weekly rankings based on nearly all tweets across Japan.

"CM Rankings"
Daily and weekly updates for CM rankings by company (covering about 3,000 companies). You can also check rankings by brand and talent.

"TV Program Rankings"
Weekly updated rankings showing which TV programs generated the most buzz.

"Trending Words (Spread Words) Ranking"
Weekly updates on words that ranked in Twitter trends. Handy.

"Trending Tweets (Spread) Ranking"

• "General Accounts" with under 10,000 followers
• "Celebrity Accounts" with 10,000 or more followers
• "Official" accounts verified by Twitter
• Tweets containing links to YouTube videos "YouTube"

We update this ranking weekly based on tweets that became trending topics across these four categories.

The ranking's accuracy is dramatically enhanced not only by using over 40,000 keywords (queries) from 3,000 companies processed by machines, but also by incorporating daily human checks. Simply viewing this ranking allows anyone to easily see what was trending that day or that week.

■Estimating Demographic Composition

We estimate which demographics are talking about trending commercials using newly developed AI. Understanding reactions by gender and age group (demographics) is crucial for gauging the success of a commercial's target appeal. Buzz Watch visualizes demographic reactions that were "expected but not available" with conventional tools.

■Positive/Negative Sentiment Analysis

We graph the positive/negative reaction levels to trending commercials and the percentage of emotions like "joy," "hate," "neutral," and "surprising." You can grasp the emotional reactions to commercials as if you could hold them in your hand. Watch out for "Negative" or "Hate" at 20% or above. At 30% or above, it often indicates a "backlash."

■Automatic Extraction of Representative Tweets

This is the core feature. What kind of tweets were representative of the buzz around the commercial? AI analyzes all tweet content and automatically estimates "representative tweets" as samples. You can view them ranked from 1st to 10th based on representativeness.

Attempting this manually would be extremely labor-intensive just to review the massive volume of tweets. Selecting representative tweets from that pool would inevitably lead to subjective choices.

By leveraging AI's ability to analyze all information uniformly, you can grasp the overall reaction without having to track every single tweet.

■Semantic Cluster Analysis

All tweets about a trending commercial are automatically sorted into semantically similar groups (≈ clusters), such as "Those who like this commercial," "Those who dislike this commercial," or "Those who like the featured talent," and visualized.

Furthermore, it automatically reveals each cluster's "positive/negative sentiment," "keywords," and "representative tweets," allowing you to delve deeply into the reality of the buzz. This is an essential feature for all corporate advertising managers.

■Fan Engagement Analysis

It automatically extracts the ratio of "new audience/repeat audience" among those who reacted to a trending commercial. Furthermore, it automatically identifies the ratio of "fan audience/anti-audience" within the repeat audience.

By combining all the features introduced so far, we've created BuzzWatch's unique metric: "Fan Engagement Level." Think of it as the loyalty level toward the commercial.

Visualizing this Fan Engagement Score over weekly trends may reveal insights previously hidden. For our special contract holders, we also provide 3D trend graphs.

For example, examining a beer company's Fan Engagement Level by gender reveals that while female support was very strong until 2018, male support surged dramatically by 2020. Since men tend to drink beer more frequently, this shift toward male support suggests it was a contributing factor to the company/brand's strong performance.

And finally, this is the feature we're most proud of for truly "scratching where it itches":

The "PowerPoint Export for Analysis Results" feature. All the data discussed so far can be downloaded in PowerPoint format. Naturally, you can share it directly with your team, upper management, or clients. With the agility befitting the dashboard era, you can use it as reporting material.

Did you enjoy exploring the world of BuzzWatch? If you're even slightly interested, we currently offer a 30-day free trial—give it a try!

Free Trial Here!

BuzzWatch Office: buzzwatch@dentsu.co.jp (We welcome all inquiries)

 

We have a place we want to go—The future BuzzWatch aims for—

 

After five years, we've reached a world where the previously opaque reality of buzz is now visible to everyone. But we still have places we want to go. Finally, let us share a glimpse of the future.

BuzzWatch has already begun its next phase and holds the potential for dramatic evolution within the next 1 to 3 years. This evolution will extend far beyond the realm of commercials and advertising alone.

For example, while BuzzWatch currently focuses on daily and weekly aggregation and analysis, the future will enable real-time visualization of all metrics. The moment a commercial airs, you'll instantly know "who reacted," "how they reacted," and "what the main comments were."

Furthermore, buzz analysis will extend beyond commercials and ads to encompass every event in society. It will enable buzz analysis for TV program reactions, trending topics on social media, and all events that stir society—such as "An earthquake struck!", "A state of emergency declared!", "A gold medal won at the sports tournament!", or "A celebrity got married!".

Checking buzz reactions will become essential for society.

And as buzz analysis evolves, the democratization of "individual voices" will advance. What this means is that the previous assumption that "each of our individual voices is too small to be heard" will crumble, making those voices easier to hear.

Even if each individual voice is small, BuzzWatch will make it possible to see the collective voice and arguments when dozens or hundreds of people with similar views gather. Governments, companies, and organizations will be able to obtain more accurate and useful feedback.

In recent years, we've also seen a phenomenon where the opinions of a few individuals become amplified through the media, significantly impacting corporate activities. If we can accurately unravel the true nature of buzz, we can consistently make sound judgments without being swayed by so-called "fake buzz."

Our development theme at BuzzWatch is " Whole-based Individuals/Points."

Previous social listening tools were all based on the "individual/point" concept, evolving to focus on local details.

However, just as most news programs begin with the "biggest events," focusing solely on your own town won't reveal the bigger picture. First, cultivate a perspective that sees the world. Next, learn to view your own areas of interest using the same metrics as that global perspective.

Once you understand the trends in how people's tweets are generated around the world (or across Japan), you should be able to see clearly why your interests "go viral" or "don't go viral."

BuzzWatch's motto is to first survey the whole landscape. Then, to examine the individual.

BuzzWatch's evolution continues.

1. Evolution to see the whole in greater detail
2. Evolution to see the details of the individual more clearly
3. Evolution to freely explore details using theme words

are currently in development. We're making it possible to see not just TV commercials, but also the buzz around TV programs and trending news stories in the world.

Why is this necessary? Because " if everyone can see the buzz, individual voices won't be fragmented." We dream of a future where we can democratize the voices of the "individual/point" that previously couldn't be heard...!

Finally, we're announcing the development themes for the next BuzzWatch. If you're interested, please contact us. Let's build the future together!

BuzzWatch Secretariat buzzwatch@dentsu.co.jp

 

 

First, " X for WHOLE." We'll continuously improve the accuracy of BuzzWatch, which focuses on commercials, using AI.

Next, " Y for NEXT ". We'll enable detailed buzz analysis for TV programs and trending keywords.

" Z for THEME ". Enables detailed buzz analysis for your chosen theme (brand names, etc.).

And finally, " X-BD BreakDown ". We're implementing cross-tabulation-like functionality for buzz analysis.

So, why not pick up this Buzz Watch that reveals the world in terrifying detail and step into a new world?

30-day free trial here!

BuzzWatch Office: buzzwatch@dentsu.co.jp (We handle all inquiries)