Last update on .

Manufacturing plants generate mountains of data throughout the day, every day. Traditionally, data has been noted on paper or analyzed in spreadsheets. However, today it can be collected automatically via sensors and analyzed with tools that far exceed spreadsheets’ capabilities. ABI Research is forecasting that in 2026, manufacturers and industrial firms will spend US$19.8 billion on data management, data analytics, and associated professional services.

For many manufacturers, there is an appreciation that operational decisions need to be based on empirical evidence rather than guesswork. The challenges are not necessarily capturing and analyzing data, rather what to analyze in the first place. Results and findings need to have a meaningful impact on operations, so manufacturers need to take a step back and devise precise objectives.

Understanding Criteria for Investing

Manufacturers are seeking optimal performance from their production and assembly lines as competitive forces squeeze margins. In addition, manufacturers hold little inventory and expect just-in-time delivery with their suppliers to accommodate smaller batches and customizations. Extracting data from production lines is integral for meeting these external expectations, as well as avoiding unscheduled downtimes. Investments in data management and data analytics are driven by the following criteria:

Cheaper Data Collection Devices: Falling sensor prices have encouraged manufacturers and suppliers to place sensors on the production lines or inside machines to collect vast quantities of information; for example, data regarding the health and performance of a piece of machinery or the production line overall.

OEM Devices Are Becoming Smarter: Machinery suppliers are now embedding sensors in their products to help clients and the companies themselves understand how a piece of machinery is performing.

Data Collection Is Enabled by Investments in Wireless Networks: The expansion in bandwidth and reduced latency delivered by wireless technologies means that increasing volumes of data can be collected and analyzed at the edge of the network or via cloud service providers. Collecting data from machinery on the factory floor is often the beginning of manufacturers’ digital transformation strategies as they try to understand their operations. Data collected can inform activities like condition-based monitoring, predictive maintenance, and the creation of digital twins.

5G Will Support Use Cases: While there is much discussion around the deployment of 5G networks, several companies and use cases are already supported by wireless networks. However, the extra bandwidth and improved latency provided by a 5G network will increase the volume of data that can be collected; for example, from tracking assets moving through the facility and/or conducting video inspection activities.

Acceptance of Cloud Services: Historically, manufacturers were reticent to use cloud service providers, preferring to retain data on-site. Improvements in security for data uplinks by service cloud providers have meant more data and workloads are performed away from the manufacturers’ networks.

Orchestration Will Be Critical: As a result of increasing volumes of data being generated, manufacturers will need assistance from software suppliers to optimize processes for data collection and storage.

Analytics Has Become More Sophisticated: Advances in AI have meant that analytics software can perform increasingly sophisticated tasks; not just reporting what has happened but anticipating events and proactively advising users of measures to take.

Recommendations for Suppliers

When introducing the IoT data analytics value chain, not only do conversations center around Return on Investment (ROI) (e.g., reducing wastage levels, increasing throughput, and/or improved quality), but manufacturers also need to consider the lost opportunities due to not investing in the value chain; savings in the short term prevent transformation, so the COI can be huge.

A successful deployment of the IoT data analytics value chain also requires clients to have a number of different skill sets. Large manufacturers will have dedicated teams for particular functions; however, smaller firms will have individuals that perform a number of different roles, so clients must be data-literate.

Suppliers need to be proactive in understanding client needs. Some are running executive workshops to explore what digital transformation really requires and how best to deliver that. At the functional level, conduct workshops with staff members to help uncover competency levels and where potential project bottlenecks might occur.

OEMs now appreciate there is immense value in the data being collected from devices in their machines and need help with strategy. Software suppliers need to work with OEMs to package the data into a Data-as-a-Service proposition. Solutions of this nature potentially could be sold to inform manufacturers how their piece of machinery compares with others in operation. Eventually, the OEM model could replicate the Information Technology (IT) market where the hardware is a loss leader and the margin is in software and services.

Both manufacturers and suppliers need to avoid pilot purgatory and “paralysis by analysis.” All stakeholders need to understand the overarching challenge the project is resolving (the why), what data is meaningful, and how the project will run end-to-end. In the context of the IoT data value chain, suppliers and manufacturers need to work on short projects and develop a momentum of delivering results quickly, and then move on to the next project. Having paid for trials concentrates the minds of both manufacturers and suppliers.

The advancements in Artificial Intelligence (AI) and machine learning mean that suppliers’ cannot just report data but also predict outcomes and suggest recommended actions. The orientation for action makes for compelling propositions, and when combined with data visualization platforms embed data in many different roles. The advent of no code/low code platforms allows staff not have to be data scientists to utilize analytics in their roles. This democratization of data means that decisions are based on facts, rather than guesswork.