A smarter, faster path to data-driven operations at scale

“Digital transformation, driven and powered by smart data, fosters new business processes and services while boosting industrial performance strategies beyond lean practices. Nevertheless, implementing such disruptive approaches effectively can present a significant challenge.” – Daniel Coudriet, Intelligent Operations Offering Director, Smart Factory COE Capgemini

Design concepts:

Be usage focused to secure user commitment: Data transformation must be driven by use cases at different levels of the company. It shall consider the business stakes and address the pain points that affect the daily life and productivity of everyone involved.

Leverage a global Industrial Data Approach to break down silos: you have to aggregate all the data related to industrial activity in order to get the right amount of information. That means collecting data from various sources: DCS¹, Data Historian, LIMS², ERP³, MES⁴, Operators, IoT⁵, etc.

• E.g. transactional and typical ERP pieces of information provide the context for analyses or information to compute benefits.

Boost agility and responsiveness through live data: If business intelligence approaches provide batch-generated reporting, industrial data relies on live streams that need to be treated in real time to provide fresh and actionable insights.

Streamlined data management: Plan for data harmonization and contextual data feeds to make raw data easy to use and then integrate those features as native services built into the acquisition/exposition stack.

Detailed design/implementation concepts:

– Consider different layers of data enrichment, adding the meta data as close to the source as possible.

– Make sure that the storage mechanisms can handle:

• Continuous, high-velocity data feeds resulting in huge volumes of data
• Contextualized and aggregated data sets based on multiple, configurable/end-data consumer dimensions: Leverage single facts (data points) with multiple attributes (dimensions), a concept also known as the “snowflake” data architecture pattern, to foster the multi-dimensional analysis you need to get useful insights.

Project management success factors:

– Consider all the aspects of the project: Consolidate the new data-driven operations mindset by ensuring technical integration, a continuous improvement culture and methodology, as well as organization and change management, training/upskilling and empowerment.

– Integrate data management as a top-level process in the organization, emphasizing reference/master data quality since that provides contextualization information and is necessary for multi-dimensional analysis:

• Consistent master data are key enablers for scale-up.
• Ensuring data quality means involving business owners, from the shop floor to the enterprise level.

– Start small (by running pilots), learn, and generalize: This is a good way to ramp up to a scale-up. It allows you to understand the pain points and usages and helps stimulate team interest. It shows the impact on the organization and makes it possible to define the change management approach for the scale-up.

– Make or buy: Consider the overall costs, factoring in implementation and benefits realization lead time and then think about how best to leverage specific skills over the commodity intelligence that is currently built into solutions.

– Coordinate OT, IT, business, and experts (data, process, etc.): Plant transformation is a multidisciplinary project that requires the commitment of various stakeholders. That is why you need to have the right project governance with enough clout to bring them all together.

– Push organizational boundaries to empower business owners to:

• Boost operations capabilities and autonomy regarding data usage instead of just relying on IT/OT teams.
• Allow autonomous production teams (APUs) to be more proactive and have more impact.
• Leverage the scaling factor of the larger business community.

Conclusion

This guidance will help you to fully and sustainably realize the benefits of data-driven operations, unlock new value by developing data-driven operating models, and take advantage of an accurate, real-time, 360° view of operations.

¹ DCS = Distributed control systems
² Laboratory information management system
³ Enterprise resource planning
⁴ Manufacturing execution system
⁵ Internet of Things

Author

Daniel Coudriet Digital Manufacturing Centre of Excellence at Capgemini Engineering Daniel is leading the Manufacturing Intelligence pillar of the Smart Factory offering, leveraging his experience with Industry 4.0 transformation projects and implementation across various sectors, from life sciences to discrete manufacturing.