Industrial IoT Must Move from Optimization to Transformation



  • Many companies have not yet fully embraced the transformational nature of the Industrial IoT (IIoT)
  • Half of IIoT projects in the past 12 months met or exceeded their goals
  • Edge and cloud computing will grow equally strong, creating a hybrid world
  • The glass is half-full, not half empty - however, the industry must accelerate the journey

Global survey unveils current strategies, success rates and obstacles

Industrial IoT (IIoT) can be seen as a way to optimize existing processes and business models, for instance by achieving higher degrees of automation, or by avoiding outages with help of predictive maintenance.

However, IIoT can and must be more than that. We have seen how new digital business models have disrupted industries like media, retail or travel and the same will happen over time to industries like manufacturing, chemicals or energy. Thus, at its core, the IIoT is not about achieving some percentage points of efficiency. It's about which companies will capture which portion of this trillion-dollar opportunity and which companies will become an extended workbench of a predominantly digital value creation.

This means that adopting the IIoT must go beyond optimization - it has to be a transformation. It means introducing new data-enabled processes in R&D, production, marketing and sales, new forms of cooperation in the supply chain, new ways of creating and commercializing products and services - all backed by a technology architecture that enables interoperability between things and provides data insights with the required speed.

Together with Industry of Things World, one of the leading IIoT conference series globally, Hewlett Packard Enterprise (HPE) conducted a survey 1 to find out to which degree company leaders approach IIoT as optimization or as a transformation, how successful they have been, and what the biggest obstacles are. We also wanted to know which technology architectures they implement how important will the public cloud be for IIoT, and which role will Edge Computing play?

Optimization goals prevail, mixed success rate of IIoT projects

Asked about the business goals they want to achieve with their IIoT initiatives, by far most respondents (64%) named increase efficiency. Similarly, other high-ranked IIoT goals like increase flexibility (48%) and reduce time to market (35%) rather aim to optimize the existing business, not create the new. In comparison, transformative goals like establishing new business models (34%), improving marketing (27%) and product development (26%) with help of IoT data, or the transformation from product sales to an as-a-services model (25%) ranked relatively low.

To be clear, increasing efficiency or time to market are important business goals however, the dominance of optimization goals in the context of IIoT can be seen as an indicator that many companies have not yet fully embraced the transformational nature of this concept.

Can we as a consequence expect that the IIoT projects of our respondents have not been entirely successful? Yes, we can. 53% said their IIoT projects in the past 12 months either met or exceeded their goals, while 47% did not reach their goals - a small portion of which even say their projects were a complete failure.

So, for which reasons did companies struggle with their IIoT projects? Respondents named the lack of skills and the culture within their own company as the biggest obstacles (both 38%). This clearly underlines the fact that being successful with IIoT requires a company transformation, a central aspect of which is that you need new skills and mindsets. This is a challenge to many companies, as well as transformational issues like missing organizational structures (27%) and wrong governance and management (21%).

The survey makes clear that transformation also applies to the technology that underpins IIoT initiatives. You can’t just buy IIoT technology – IIoT requires a fundamental redesign of the information technology (IT) and operational technology (OT) architecture.

You can't just buy IIoT technology you have to transform your technology architecture

Closely after skills and culture, missing standards (36%) is named as one of the biggest IIoT roadblocks. This refers to the difficulty of making the “things” in the IIoT speak with each other – be it a Kuka robot speaking with an ABB robot, a production machine speaking with an IT systems, or a manufacturing plant speaking with an energy supplier. This requires common standards as well as new technology architectures which create convergence of information technology (IT) and operational technology (OT).

Accordingly, the number one skill companies are looking for is the ability to design new common architectures for IT and OT (45%), and number three is the ability to create a unified approach to the operations and support of IT and OT (32%). The fact that software development is also one of the most desired skills (42%) shows that we need new types of IT/OT software which in many cases are not available on the market as a packaged application.

Edge and cloud computing will grow equally strong

We also asked which role edge and cloud computing will play for IT/OT architectures in the coming years. Many market observers have emphasized the crucial role of the cloud for the IIoT. However, as our survey shows, edge computing is as important and will grow equally strong.

Edge computing means that compute and analytics resources are not running in a central datacenter or a cloud, but near to the things or in the things themselves - think of the IT embedded in a car, IT systems embedded in a machine, running on the factory floor or on an oil rig.

First, if we look at how much of the sensor data is today analyzed on IT systems close to the data source (i.e. the edge) and in the cloud, the percentages are quite low on both ends, with more than half of companies analyzing less than 30% of sensor data at the edge and in the cloud. The rest of the sensor data is processed in a traditional company datacenter.

But if we look five years ahead, these numbers will change significantly, the ratios will be turned on their heads, with the majority of companies planning to analyze 30% to 70%, or even more than 70% of the sensor data at the edge and in the cloud.

This might seem to be a contradictory result, but it is not. In order to realize the promise of the IIoT, we have to deeply transform our traditional IT architectures, and we need both the edge and the cloud capabilities to make that happen.

Top-three reasons for edge computing: security, latency, bandwidth

Why is this the case? We also asked that question. First, the three most important reasons for using edge computing in the IIoT are security (52%), latency (41%) and bandwidth (35%). Let’s look at latency and bandwidth first. Imagine a self-driving car driving with 100 kilometers per hour towards an obstacle on the road – the IT systems in the car have to analyze mega- and gigabytes of sensor data within milliseconds to avoid a crash. There is simply no time to send that sensor data to a remote cloud and wait for an answer. This equally applies to production machines and other things. And security: sending all that data via the network opens a big attack vector for hackers, so it’s better to analyze the data on site and only send selected and encrypted data to the cloud.

Similarly, there are important reasons to use the cloud, the top three being correlation analysis (66%), deep learning (51%) and horizontal integration (36%). It’s not enough to have intelligence in one machine, car or plant – you can create more value if you bring the data of these machines, cars or plants to one central place to be able to compare and correlate their behavior. Then you are able to derive deep insights from the data – i.e. do deep learning – which you can play back to the things and enable them to perform better, adapt to new and unknown situations better, and avoid outages. This also allows us to better coordinate cars, machines and plants, enabling things like swarm intelligence in traffic or highly automated supply chains.

This means the IIoT will be a hybrid world. And one of the key tasks will be to create integrated architectures that bridge from the edge to the core datacenters to the cloud and all the way back.

Encouraging results, but still a lot of work to do

Again, the journey towards the IIoT goes beyond optimization - it is a transformation which requires change on all company levels: technology, architecture, processes, people, and business models. Overall, our survey results show that the industry is still in the learning curve in that regard. However, we have to consider that IIoT is an emerging concept, and therefore its encouraging to see that transformational approaches already play a significant role in the way companies plan and execute IIoT. Similarly, Id suggest to talk about a 53% success rate, not a 47% failure rate. The glass is half-full, not half-empty. This also means theres still a lot of work to do. My appeal to the industry is: Embrace transformation and accelerate your journey, because theres not much time.


Learn more at:

1. State of the Industrial IoT, September 2017, based on surveys with 350 managers, directors and C-levels between July 2017 and September 2017. 68 percent of respondents were from Western, Central and Eastern Europe, 14 percent from North America, and 13 percent from Asia/Pacific. Most respondents were participants of one of the Industry of Things World conferences, indicating an above-average IIoT maturity.


Adapting and Innovating with the Internet of Things

Blog Post

From Steel City to Smart City

Blog Post

Staying on the Cutting Edge with Software Defined OT

Blog Post