To prevent big data from becoming a wasted treasure. What are the secrets to successfully leveraging a "Data Warehouse (DWH)"? (Part 2)

• To prevent big data from becoming a wasted treasure. What are the secrets to successfully utilizing a "DWH (Data Warehouse)"? (Part 1)

In recent years, we often see many companies pushing forward with business digitalization and creating business plans centered around keywords like "DX" and "big data." However, it seems many companies are struggling, saying, "Even if we just set up the framework and structure, we don't know how to actually apply it to our operations."

Continuing from Part 1, we spoke with Shinichi Mienaga of Dentsu Digital Inc. about corporate challenges surrounding data utilization and the benefits of implementing systems like a Data Warehouse (DWH).

Analyzing accumulated data based on defined challenges is fundamental

Q. What benefits can be gained by collecting and analyzing big data using a DWH—a database that consolidates and stores the vast amounts of data accumulated daily in business operations in a chronological manner?

Yutaka: That depends on the purpose—whether it's solving marketing challenges or sales challenges. Taking the pharmaceutical industry I work in as an example, a key challenge for drug companies is determining what sales activities, when conducted by representatives meeting physicians, increase the likelihood of their drugs being considered for prescription.

In this case, effectively utilizing data can reveal insights like "bringing this type of material increases the likelihood of closing a deal." This data includes, for example, accumulated records showing which materials past sales reps brought to clients that ultimately led to orders. Properly analyzing this data helps solve sales challenges.

Q. So, there are several examples like this among the projects you've handled so far, Mr. Yuteki?

Yuteki: That's correct. Through working on numerous projects, we've become able to clearly communicate things like, "In this industry, collecting this type of data is effective," or "If you have this particular challenge, please prepare this kind of data."

That said, defining the problem is essential for utilizing data. We must never forget that data collection and analysis exist to solve that problem. I believe the most important thing is to propose the right solution by comparing the desired goal with the current situation.

Q. Even if a company sets the challenge, prepares the environment, and implements the solution, will some companies successfully utilize the data while others struggle?

Beneficial: That does happen. Companies with dedicated digital marketing departments that drive projects effectively tend to utilize data well. However, if responsibilities and accountability aren't clearly defined from the start, it becomes difficult to get things on track. Additionally, companies with frequent job rotations need to be cautious. Even if they implement a system, if the responsible person leaves after a year, effective utilization often fades.

AI analyzes detailed customer information to enhance sales precision

Q. You're involved in many AI-related projects, Mr. Yuteki. Are you seeing more projects linking collected data with AI?

Beneficial: Yes. AI-related projects have increased significantly recently. Take a certain automaker, for example. They had already implemented a data warehouse (DWH) and were utilizing data to a reasonable extent. However, their dealers were manually creating prospect lists, and approaches based on those lists weren't very effective.

They approached us asking, "Can we use AI to fix this?" But typical AI applications only predict conversion rates for potential customers. That wouldn't solve their problem.

Our approach was to use XAI (Explainable AI). This provides actionable insights for customer outreach, such as the customer's past vehicle replacement intervals, preferences, time remaining until the next vehicle inspection, and current mileage. This enables the client to use their prospect lists more effectively.

Companies with different data collaborate to create new value

Q. Companies hold data that may include personal information. They often ask us to consult on how to utilize this data for marketing, even though they can't disclose its contents. How do you handle such cases?

Beneficial: In such cases, we often employ statistical processing on the data that can be disclosed. We then deliver it in a format that can be combined with the non-disclosable data held by the client. This approach focuses more on refining and enhancing the data's utility rather than utilizing the data itself. Alternatively, we sometimes partner with companies possessing data at a level that can be disclosed to conduct joint projects.

For example, it's crucial for insurance companies to identify potential customers with future disease risks and target them appropriately with relevant insurance products. However, their own data alone doesn't reveal which attributes correlate with higher disease risks. Therefore, they collaborate with companies that hold data on "people who visited a hospital for a medical examination."

While information like "who developed this specific illness" cannot be disclosed, analysis can be performed using attributes such as residential area, age, and gender to determine "people with these attributes are more prone to certain illnesses." Providing this information to life insurance companies allows them to combine it with their own customer data. This enables them to tell specific customers, "People with these attributes are more likely to develop this illness." This becomes a persuasive argument for customer acquisition.

Q. So even if there are some restrictions on the data available for analysis, it can still be effectively utilized by integrating it with other data, right?

Beneficial: In the past, there was a strong tendency to collect vast amounts of data directly from customers within the company and base analysis on that. However, I believe there is now a growing trend to purchase or collaborate with data held by other companies and effectively combine it with the company's own data for utilization.

Q. Finally, what other possibilities do you see for future data utilization?

Beneficial: Well, for instance, linking data to specific individuals could be a possibility. Whether in sales or marketing, people remain a significant factor. What if we could create a system that suggests, based on data, which internal team members would work well together on a given project? By accumulating information internally like, "This person achieved results in this field with this type of proposal in the past," it could lead to decisions like, "Then, for this new project, let's assign this person and that person." I believe this would increase the likelihood of project success.

The realm of data utilization, including DWH, still holds immense potential. It seems a tremendous waste to pursue DX merely as a facade, driven by reasons like "it's trendy" or "it's an executive request." By taking the proper steps, clarifying challenges, and defining clear goals, the truly useful, unique form of DX tailored to your company will naturally emerge. Why not use DX promotion as an opportunity to re-examine the future direction of your business?