Big Data from an Advertising Agency's Perspective

"Is ××× an enemy or an ally to our company?"
This is a question frequently heard within our company, used to broadly assess the nature of unknown phenomena that appear before us. And right now, 'big data' is surely one such phenomenon for which not only Dentsu Inc., but the entire advertising industry, is searching for an answer.
This column will primarily explore the theme "What is big data for an advertising agency?"
"What is it? You mean combining media exposure data, social data, and purchase data, then modeling it using advanced statistical techniques? We're already working with big data at our company,"
That's exactly right.

That's correct.
Currently, within the advertising industry, efforts to leverage big data—analyzing comprehensive data like behavioral logs as an extension of traditional panel-based data analysis, primarily for media investment accountability—are intensifying and becoming a competitive arena.

Within this context, this column aims to explore new business domains and marketing strategies that advertising agencies should pursue regarding big data, introducing examples that could serve as seeds for their realization.

It's not the large volume that defines it, but three characteristics that "result in" large volume.

First, what business domains should advertising agencies aim for by leveraging big data?
As the Ministry of Internal Affairs and Communications defines big data business as "businesses that use big data to solve social and economic problems, enhance the added value of operations, or support such activities" (※1), it could also be described as "businesses that enhance the added value of client operations or support such activities."
Big data itself emerged suddenly about three years ago. It's rare for a buzzword whose true meaning remains unclear to gain such widespread acceptance. However, we cannot proceed with an ambiguous understanding of big data. Therefore, this column will base its discussion on the three characteristics introduced by the Ministry of Internal Affairs and Communications: "high resolution," "high frequency," and "diversity." Essentially, these three characteristics mean big data may "result in" large volumes, but "large volume" itself is not an essential defining feature.
Misunderstanding this leads to a focus solely on the technology for processing large-scale data. Before you know it, large-scale data processing systems like Hadoop(※2) are built within the company, and only later does the discussion begin: "What business value can we actually derive from this big data?" This is putting the cart before the horse.

The potential for sensor-derived data to drive new business and marketing innovation

What specific data can help advertising agencies enhance their added value?
Personally, I expect sensor-derived data to be the catalyst for new business ventures and marketing innovation. Alongside smartphones, "network-connected sensor technology"—including wearables—is becoming widely adopted. The photo below shows my "wireless scale." Stepping on this scale automatically uploads diverse data—like my daily weight and body fat percentage—to the cloud at high frequency. I can then track these numerical changes across various devices.

By leveraging big data about individuals—collected from sensors and uploaded to the cloud—with the individual's consent, service providers can collaborate and utilize this data to offer more compelling services. For example, healthcare providers can use patient motion data and sleep state data to detect abnormalities in real time, enabling the delivery of higher-quality medical services.
As the broader perspective of marketing shifts toward "real-time & personalized," CRM management using sensor-derived data holds significant potential for us as advertising agencies.

Next time, I'd like to discuss the potential of sensor data utilization from a marketing perspective.

(※1) Ministry of Internal Affairs and Communications, Information and Communications White Paper 2012 Edition: What is Big Data?
(※2) Hadoop: Open-source middleware for efficiently distributing and managing large-scale data processing.