Published on 02/21/2017 | Market Sizing
1. Status quo - Firms acknowledge the huge importance but are not yet completely set-up
• The importance of analytics for decision-making is increasing: Analytics started as mere operational support in the 1960s and 1970s. Today, it is increasingly used to drive decision-making. In the future, it will be used to automate decisions.
• 15% of respondents surveyed view industrial data analytics as a crucial factor for business success today, while 69% think it will be crucial in 5 years.
• Today, 68% of survey participants say they have a company-wide data analytics strategy, 46% have a dedicated organizational unit and only 30% have completed actual projects.
2. Value drivers – Increasing revenue seen as the main driver; predictive maintenance as the leading application
• People see increased revenue as the main value driver for Industrial Analytics (33% weighted score). Increased revenue can be achieved in three possible ways: Upgrading existing products, changing the business model of existing products, or creating new business models.
• Despite the fact that one can witness a number of e ciency-related projects today, cost cutting is seen as less important at only 3% (weighted score).
• The three main applications of Industrial Analytics in the coming 1-3 years are related to predictive and prescriptive maintenance of machines (79% of respondents view it as important ), customer/marketing- related analytics (77%) as well as the analysis of product usage in the eld (76%).
3. Analytics – Slowly shifting to more sophisticated types of analytics
• The type of analytics deployed on various projects are moving from descriptive analytics to applications of real-time analytics, predictive analytics and even prescriptive analytics.
• The importance of spreadsheets will decline (from 54% to 27% in 5 years) while the importance of Business Intelligence (39% to 77%) and advanced analytics tools (50% to 79%) will increase sharply.
• IoT brings additional challenges to Industrial Analytics, including real-time data streaming, management of enormously large data sets, time-stamp data storage and completely new use cases –Most companies feel they are good or excellent at collecting IoT-related sensor data (60% of survey respondents) but only few say they are good or excellent at getting the right insights from the sensor data (32%).
4. Paradigm shifts – Industrial Analytics changes long-held manufacturing principles
• Agile project development is replacing waterfall-based project planning. 58% of survey respondents indicate that they employ the agile (and often also “scrum”) methodology for their data analytics projects today.
• Other paradigm shifts include the creation of platforms and open ecosystems (e.g., companies are building B2B marketplaces and app stores”), the reshaping of the well-established 5-layer automation pyramid (software architecture), as well as an increasing exibility and specialization of manufacturing through manufacturing-as-a-service.
1. Starting the project – Often in an explorative approach and using open source tools
• In their quest to embrace digital business models and build on the power of data, companies start projects increasingly in an explorative manner (34% use an explorative approach) – still, the majority (66%) of projects are approached with clear hypotheses in mind (hypotheses driven approach)
• 4 areas need to be addressed, when structuring Industrial Analytics project: Data sources, necessary infrastructure, analytics tools and applications
• Using open-source analytics tools are increasingly the norm: Nearly two thirds of all survey respondents (64%) are using open-source tools for some aspects of their data analytics projects.
• Most costs in Industrial Analytics projects incur in the initial phase of getting data access (21%), aggregating the data (17%), and performing the data analysis (14%) – the costliest individual item, however, is related to software and application development (26%).
2. Organizing and Staffing – Top management-driven, externally implemented – bridging the Data Science Skill Gap
• Industrial Analytics is increasingly initiated by senior management - 34% of survey respondents indicate that it is the CEO who drives Industrial Analytics projects.
• Large corporations have not centralized data analytics in one speci c department (Only 33%). Instead, many large industrial companies are outsourcing some of their data analytics activities in an external Data lab, Digital lab, incubator or accelerator (55% of respondents)
• The biggest skill gap is currently in Data Science. (92% of respondents say it is important or very important but only 22% of respondents have all necessary skills on board). Machine Learning, as an integral part of Data Science also represents a large gap (83% vs 33%) – Another signi cant de ciency can be identi ed around IoT/M2M infrastructure (68% vs 17%).
• Data Science Teams are diverse and typically include an overall manager, an industrial expert, a data engineer, a data developer, a Machine Learning expert and a data analyst.
3. Challenges & further Recommendations – Focus on interoperability issues, data accuracy and shaping the digital mindset
• Overlapping tasks with departments (60%) and di culties in building the business case (60%) represent the most important business challenges for IA Projects
• Interoperability between di erent components of the data analytics IT/OT stack (78%), data accuracy (62%) and gaining insights from data (62%) represent the biggest technical challenges
• Further leadership recommendations: Shape the digital mindset, de ne strategic roles, start small, de ne a capability roadmap, embrace a data governance strategy, and enable supporting functions.
You can download the full report here.
