[Survey on Data Integration in Manufacturing] Approx. 70% Operate in Environments Including On-Premise, 60% Retain Manual Integration Using Excel/CSV

Three-shake conducted a survey on the state of data integration in the manufacturing industry, targeting 100 professionals across Japan.

The results reveal that approximately 60% of companies perform operations involving manual work using Excel or CSV, indicating that data integration still heavily relies on human intervention. Furthermore, about 70% of respondents stated they operate in configurations that include on-premise environments—either as a mix of on-premise and cloud or as on-premise-centric setups—highlighting challenges such as employee burden and dependency on specific personnel within existing system environments.

■ Background
In recent years, interest in DX (Digital Transformation) and data utilization has grown in the manufacturing industry. However, in actual operations, long-used core systems, various SaaS, and Excel-based workflows are intermingled, making cross-departmental data integration and utilization a significant challenge. Particularly in manufacturing, many cases involve maintaining existing on-premise environments while introducing cloud services, which tends to complicate system configurations.

Given this backdrop, Three-shake conducted a survey to investigate the realities of data integration in the manufacturing sector.

■ Survey Overview
Survey Name: Survey on Data Integration in Manufacturing
Method: Internet survey using the self-research service "Sacrisa"
Duration: May 2026
Respondents: 100 individuals

■ Survey Results
・Approx. 70% operate in "environments including on-premise"
41.2% reported a mix of on-premise and cloud environments, while 33.3% said on-premise is central, confirming that about 70% of companies include on-premise configurations.

・Data integration method: "Partial automation + manual work" is the most common
The most frequent response regarding data integration methods between systems was "partially automated but includes manual work (such as CSV output)," at 36.3%, followed by "center on manual work" at 30.4%.

・Challenges include "manual work burden" and "modification costs"
Regarding current challenges, "high volume of manual processing" accounted for 23%, and "time/cost required for modifications or changes" accounted for 20%.

・Approx. 50% feel challenges in efficiency
Combining "hardly streamlined" (22.5%) and "not very streamlined" (29.4%), about 50% perceive challenges in efficiency.

・Approx. 50% report "dependency on specific personnel (siloing)"
Regarding siloing of data integration tasks, 21% said they "strongly depend" on specific individuals, and 33% said they "partially depend."

・Approx. 40% perceive a "need for improvement but have not started"
Regarding future plans, 11.8% said "it is necessary but have not started," and 27.5% said "are considering it," showing about 40% recognize the need for improvement.

■ Conclusion
The survey clarifies that while DX is progressing, significant challenges remain in the underlying systems that connect data. The fragmentation of systems is leading to increased modification costs and dependency on specific individuals.