The Intersection of Analytics and the Internet of Things

Aapo Markkanen and Dan Shey
November 11, 2014


The business case behind most Internet of Things (IoT) deployments relies on collecting data and gaining actionable insight from them through the right types of analytic tools. Mere connectivity already allows valuable enhancements such as remote service, but ultimately the value in IoT can be found in the ability to expose detailed and comprehensive (yes, even "big") data from assets and processes that have traditionally been more or less opaque to accurate analysis.

By taking advantage of this crossover – which we refer to as IoT analytics – the connected organizations can build their operations on hard evidence and statistical probabilities instead of soft opinions and gut feeling. Or that is the goal at least. It should also be borne in mind that a data-driven organization can be driving in a very wrong direction if the insights that it has gained are incomplete or outright misguided. It is a cliché to remind that correlation does not necessarily imply causation, but in such a perfect storm of buzzwords it is still a point worth making.

Now that we are done with that disclaimer, let us take a look at what this intersection is actually made of, starting from the value chain that enables the whole concept. ABI Research segments the analytics market into five different components, which are listed and described in the following.

Data Integration: Aggregation and integration of the collected data streams in a manner that makes them suitable for analysis. This tends to be more critical in IoT-driven fields than in those that are digital by design, as end-users normally must deal with more diverse and disparate data sources before meaningful analysis is viable.

Data Storage: Implementation and management of the data store holding the data sets that have been integrated for the analytics process. In IoT, the key issue to address in this context is how to store time-series sensor data, which can increase dramatically in volume compared to e.g., transactional data readings.

Core Analytics: Processing of the data by an analytics engine and the subsequent delivery of insights – covering all of the so-called three phases of data analytics: descriptive, predictive, and prescriptive. In the case of IoT, the evolution from the descriptive phase to the predictive one is currently three years or so behind what is being seen in the "digital-first" industries.

Data Presentation: Further presentation of the delivered analytical insights to the end-user in the form of reports, visualizations, or dashboard mash-ups. In IoT analytics, the geography of data is a particularly important presentation element, given that the location of physical Things matters more to analysis than, say, the location of an ecommerce transaction.

Encompassing all of the four technology components is the fifth segment of the value chain: Professional Services. This segment refers to the services provided by various types of external consultants, either on a one-off or ongoing basis, to facilitate the process.

And what does this value chain actually enable in practice? Based on the conversations that we have had with organizations that have been early movers in this space, the main use cases – roughly speaking – can be generalized under five distinct categories.

Predictive Maintenance: Predictive maintenance refers to a method in which equipment or infrastructure is maintained when an analysis of its operational (e.g., sound, speed, vibration) metrics indicate that a breakdown is likely to occur. The condition-based method can be complemented by circumstantial data (e.g., ambient temperature, employee absences, product recalls) to make the analysis more accurate. Considering that maintenance is a characteristically labor-intensive activity, optimizing it in this manner can bring substantial cost savings.

Product and Service Development: As an analytics use case, product/service development aims at assessing the connected product's quality and behavior, and then flagging up areas of improvement based on the assessment. For instance, a tractor manufacturer could study the real-life usage of its latest model to not only conduct predictive maintenance, but base the following iterations' modifications on the analysis of how the customers tend to operate the machines.

Usage Behavior Tracking: Usage behavior tracking (or usage incentivization) refers to cases in which usage or consumption of a product or service are tracked and analyzed by taking advantage of IoT connectivity and subsequent analysis of the collected data. Analysis of, say, electricity consumption within a smart grid could be used to reveal high-usage customers, and then mitigate their consumption through targeted efficiency programs. Also car insurance providers and other companies applying usage-based pricing to their offerings are usually counted under this category.

Operational Analysis: In operational analysis, the organization employing IoT analytics applies the data assets to monitor and optimize its internal operations. Notably, many of the applications in the transport and logistics segment can be best characterized as a form of operational analysis. For instance, a logistics group can analyze its delivery fleet to optimize routes and provide more accurate estimates on delivery times. Similarly, a retailer running connected vending machines can spot the bottlenecks and quiet zones within its network, and thereby optimize the machine sites.

Contextual Awareness: In this use case, a connected object collects data from the surrounding dynamic environment and adjusts its operation accordingly. The gained contextual awareness can be then used to similarly "smarten up" other objects within the same network. Concepts related to assisted or autonomous driving and various forms of advanced robotics are a prime example of IoT plays that rely on contextual awareness. Similarly, a connected thermostat can adjust its timer settings in response to the home owner leaving from work earlier than usual, or in preparation for a forecasted arrival of colder-than-expected weather.

In essence, what all these use cases have in common is that they are, in different ways, making the physical world more digital and thus more transparent for decision making. More transparent premises for decision making, in turn, allow more informed decisions. This is particularly transformative because changes in, and interactions with, the physical world are by definition much more far-reaching and irrevocable than the ones that take place exclusively in the digital domain. At the end of the day, that fact is also what makes the said intersection so intriguing. It represents a promise of less wasteful and more sustainable decisions, in very many parts of business and society.



Aapo MarkkanenPrincipal analyst Aapo Markkanen leads ABI Research's Internet of Everything Research Service, contributing to various research activities related to Internet of Things, M2M, and big data. In his research, he explores areas such as predictive analytics, product lifecycle, quantified self, contextual awareness, cloud platforms, and IoT developers. Before joining ABI Research, Aapo worked as an analyst at IHS, where he was responsible for providing market intelligence and strategic analysis on the European telecoms sector and its leading players. He holds BSc and MSc degrees in management studies from the University of Tampere, Finland.


Dan SheyPractice director Dan Shey manages ABI Research's M2M/IoT and enterprise mobility research services covering the telecom and IT ecosystems with a focus on devices, applications, convergence, and strategic analysis of the industry's value chain. Prior to joining ABI Research, Dan worked as an independent technology business consultant. Earlier still, he worked in product management, product development and marketing at Qwest Wireless. Dan holds a BS in Physics from Loras College and an MS in Metallurgy from Iowa State University. In 2000, he received his MBA from the University of Michigan where he is also a Fellow of the school's Tauber Manufacturing Institute.