A group of industry experts assembled at LiveWorx last week to discuss big data’s relationship to the Internet of Things.
How organizations are monetizing IoT’s big data
The IoT is defined by massive networks of connected products, all creating and sharing information with each other. In one sense, data can be viewed as the byproduct of IoT activity, but that sells short the value of big data.
IoT data represents a whole new class of information: quantifiable intelligence about the operation and performance of products after they’ve left the factory and gone into use.
This data can be used to better service existing models, improve how new models are built, improve customer relationships, optimize billing, accelerate engineering, and even reassess the entire design and purpose behind products.
Big data’s IoT service maturity curve
Service is one the most immediate and substantial ways that an organization can benefit from IoT data. The usage of IoT big data for service follows a maturity curve, similar to PTC’s maturity model for smart, connected products. The curve plots increasingly ambitious service stages, each requiring IoT-focused infrastructure. The curve also requires businesses to trust data-driven systems.
These service stages are as follows:
Descriptive. Collected data is presented to operators or supervisors. Data presents accurate, real-time, quantifiable benchmarks about how products are operating and performing. If assets are failing or operating below optimal conditions, this can be observed remotely.
This requires being able to collect, filter and present data back in a way that is intuitive to operators, including notification triggers. Descriptive capabilities can accelerate service and support engagements.
Diagnostic. Collected data is collected and aggregated over time, to present more meaningful information to operators or supervisors. Individual data points, collected in the descriptive stage, provide visibility to specific, single attributes.
On their own, temperature, RPMs, pressure, etc., are useful inputs for understanding one aspect of the product. But by analyzing multiple inputs, additional conclusions can be drawn, and diagnostics is enabled. Is a product momentarily stressed due to external conditions, or is there a true operational problem?
This stage requires analytics to measure and combine multiple data inputs over time, and to form diagnostic conclusions against on previous performance. Diagnostic capabilities further accelerate service engagements, and enable some self-service.
Predictive. The next phase is to take historical aggregated data used for diagnostics, and to further analyze it and apply some basic artificial intelligence and rules for identifying triggers.
By mapping multiple inputs to historical product operations, it is possible to identify the warning signs of future operational problems. This stage requires the capabilities of the previous stages, with greater analytics and some intelligence to make the leap from diagnosing past performance to drawing conclusions about future performance. Predictive capabilities can be used to take actions that mitigate the need for service, and extend product life.
Preventative. The most advanced stage of big data-driven IoT service is preventative capabilities. This requires all of the capabilities of the previous stages, while introducing manual and automated responses.
The preventative stage takes the predictive data, and maps it to response actions. Preventative applications actuate direct changes to products (e.g. lowering RPMs, reducing throughput, increasing temperature, etc.) to prevent further status changes that could lead to product stress and failure.
This requires more advanced analytics and intelligence. More importantly, it can require a change in organizational thinking, by placing trust in automated systems and big data analytics to take greater responsibility in ensuring uptime. If correctly implemented, the gains can be huge, with dramatic decreases in downtime.
The panel experts arrived at their assessment from different touch points in the IoT revolution, but agreed that IoT success, and the larger trajectory of technology, depends on effectively capturing information, securing it, and being able to turn into something actionable.
While much of the emphasis on the Internet of Things is aimed squarely at how connected products will improve the lives of consumers and operators, big data experts posit that products will actually help improve themselves.