During my last visit at IMD Business School in Fall last year, I spoke with a program advisor, who informed me about a new course within the Digital Transformation curriculum I am currently attending to: an Internet of Things (IoT) module was being prepared. I happily registered and am now sharing my learnings and reflections around IoT.
IoT is a hype topic, as much as AI or Blockchain, but it is often neglected by the mainstream press. However – and probably because of my professional background in the telecommunication industry – this new tech appears to me as the most promising one, as it represents the foundation layer for many innovative use cases, being industrial or consumer-based.
“As digital sensors and connectivity become embedded in more and more objects, the Internet of Things is likely to redefine the boundaries of many industries that were less transformed by the Internet than were media and information business.”
David L. Rogers, The Digital Transformation Playbook
Learnings & Take-aways
#1 “Design Thinking is a Must-First-Step”
Design Thinking is a methodology, structured as a non-linear creative process. It changes the paradigm of the design activities, becoming strategic (initiating the product or service development) instead of tactical (at the very end of the requirement gathering).
Design Thinking helps people and organizations cut through complexity. It is great for innovation, and works extremely well for imagining the future – yet it is not the right set of tools for optimizing, streamlining, or otherwise operating a stable business.
“Rather than asking designers to make an already developed idea more attractive to consumers, companies are asking them to create ideas that better meet consumers’ needs and desires.”
Tim Brown, Design Thinking, HBR
By iteratively developing an IoT prototype, and confronting it to customers’ actual use, one can learn about the strengths and weaknesses of the idea, and identify new directions that further prototypes might take (pivot).
#2 “IoT is Much More Than a Connected Device”
The current Top 3 categories of Internet of Things use cases are Smart Homes, Wearables, and Smart Cities. Interestingly enough, all very much consumer-oriented. But the future of IoT seems to actually lie around B2B applications, such as Preventive & Predictive Maintenance, Asset Tracking, Smart Metering, Automatic Refillment, etc.
Providing simple connectivity and remote access to devices has been around for the past 20 years, but collecting the data – and making actual use of them – is the real added value for modern IoT products.
They profit from better technology, of course (better sensors, more efficient power management, more connectivity choices), but the related use cases are highly augmented by the additional computing infrastructure (cloud & edge) and the state-of-the-art analysis capabilities (artificial intelligence).
“The key insight from a management perspective is that the source of innovation does not lie within a single technology; it is the fusion of different technologies that drives innovative IoT solutions.”
IEEE The IoT and Digital Transformation: Toward the Data-Driven Enterprise
#3 “Machine Learning is the Game Changer for IoT”
Linked to sensors and their associated data, Artificial Intelligence, and especially Machine Learning, leads to novel applications: Large-scale data collection for informed decision-making, Virtual twins, Facility maintenance & infrastructure repair, Personal assistants, etc. – the list is growing every day.
Big Data analytics for real time monitoring and diagnosis appeared a little bit, and AI came slowly along, to make better, holistic business decisions. Programmed rules & processes are nowadays replaced by trainable systems which “learn” by themselves, thanks to the vast amount of existing data.
This concept, called Machine Learning, is a subset of Artificial Intelligence, yet already the most promising one. Reinforced learning (reward system for AI agent based on effect of decisions, similar to a child’s way of learning) and Deep Neural Networks (competing or collaborating sets of AIs) are trending methods. Interpretable or explainable AI (making sense to a human being of how neural networks act) will be the next use case.
Machine Learning is not limited to IoT, on the contrary. But its specific association with connected devices enables the following benefits:
- Boost operational efficiency and prevent unplanned downtime
- Better manage risks and opportunities
- Trigger new / enhanced products and services
- Increase IoT scalability and optimize data collection
#4 “The Main Challenges of IoT are Security & Ethics”
A compromised IoT device can result not only in security breaches on the data level, but could also tamper the very function of the device. Think about hacked autonomous cars or malfunctioning life-depending healthcare devices!
A consequence is that Vertical markets are formed in IoT, where companies own and develop the different components (Hardware + Cloud + Data). Google, Amazon, or Microsoft propose “off-the-shelf” IoT solutions already.
This is a little bit counter-intuitive as higher margins are expected in Data (analytics, machine learning), but controlling the whole chain helps securing data (complete solution) and are appealing to customers – even if it comes with a higher price!
Due to the very nature of unsupervised learning, the applications of IoT have a huge impact on legal aspects. As there is no prior programming by a human being and no algorithm as such, who bears the responsibility in case of an incident?
And even though a human is still making the final decision, the computed recommendation is generated based on input data which may or may not be biased, unbalanced, altered by local or social specifics, etc.
It is not only the classical Moral Machine paradox (“should the autonomous car hit the elderly or the infant”) but it can go further on with more positive scenarios: for example, could a self-driving car drive above speed limit in case of emergency? But if so, how is the urgency communicated, detected, and evaluated by the car’s AI?
Ethical and security concerns must in many cases be considered before a system is deployed. It is therefore critical to ask those questions during the design phase, and even consider exotic skillsets such as social engineers or psychology experts.
#5 “Network Effects are at the Heart of IoT”
Metcalfe’s Law states that the value of a network increases to the square of the number of components. This exponential growth is counter-intuitive to our linear approach of change, but this actually underlines the whole digital transformation paradigm, from Big Data to the platform business models (“The Winner Takes All”).
We will soon have 50 billion IoT devices – actual small computers – connected together as a network. Fifty billion squared is equivalent to the number of stars in our universe. This powerful global network becomes a new computing platform!
And much of the computing will take place within the sensors at the periphery of the network rather than at the core of the network (edge computing).
Combine this statement with Moore’s Law (roughly stating that the computing power doubles every two years, consequently that while capability steadily increases, the cost of those capabilities continually drops), and you can feel the sheer scale of the IoT revolution.
Technology becomes faster, smaller, and cheaper. The day is coming when it will be technologically feasible, and economically viable, to network-enable even the simplest of things, right down to the light bulb.
Further Reflections & Opportunities
Project Management & Design Thinking
Design Thinking sees the big picture: an innovation is not an end, the actual value is coming from how this innovation relates to the overall customer/consumer journey.
Therefore, to be efficient and effective, it requires teams that are diverse, covering various functions, and having broad skill sets.
This is very similar to a project team, especially under an agile approach. The many similarities with the Agile project management methodology are:
- Both Design Thinking and Agile deal with cross-functional, multi-disciplinary teams;
- They are a method/process and a mindset:
- They both have iterative & customer-centric approaches (prototypes vs. MVP).
Stakeholder Management in Project Management and Avatars/Personas in Design Thinking are very similar activities; that said, Personas may contain more information than the anonymized stakeholders used in project management mapping techniques.
Opportunity: use “personalized” stakeholders (with a faked background and an emphasis on emotions and feelings, etc.) during the stakeholder analysis in order to bring additional insights.
Design Thinking in a B2B Setup
How does Design Thinking look like in B2B setups? What is a “customer journey” in this case? Who should you empathize with when dealing with a firm? These questions kept popping into my mind as I went through the learning material.
Indeed, Design Thinking is a very powerful tool when it comes to depict a typical end user, which can be categorized by demographics, behaviours, interests, power, etc.
In a B2B environment, context plays a bigger role. Additionally, we are not dealing with only one customer – different viewpoints (usage, feasibility, financial, …) must be taken into account.
This is actually why Stakeholder Management is important here, but designers struggle to capture the Big Picture.
“Design for business-to-business is not pretty. (…) Many B2B solutions are satisfactory for managers but useless or unusable for employees.”
Masakazu Iwabu, What complicates B2B design?
Opportunity: consider the functions involved in dealing with an IoT product or service (purchaser, operator, end user, developer, etc.), and combine stakeholder mapping with Persona-driven outcomes.
Project Management & AI
With regards to Artificial Intelligence, I already figured out and integrated in my career consideration that AIs are going to replace project managers sooner or later – at least for certain tasks.
The main idea is to use AI to predict project behavior thanks to typical project parameters (costs, duration, etc.) and historical data from past projects. For example, occurrence of red flags, cost overruns, or delays could be predicted – and hopefully prevented – if early signs are met.
Alerts and a possible way forward could be provided in real time to sponsors without any active involvement of the project manager, enabling transparency and informed decision-making for project stakeholders.
Somehow, this shall render the PMO function redundant, not necessarily the project manager role itself. Indeed, a project manager’s added value lies in the soft skills and in the understanding of the project context (holistic approach). Mandates and job roles requiring solely technical project management competencies (reporting, planning, organizing) shall be avoided as their future is clearly endangered.
“Machine learning experts, for their part, need to realize the gap between cutting-edge science and organizations’ ability to actually implement working models aimed at real problems.”
Getting Value From Machine Learning Isn’t About Fancier Algorithms – It’s About Making It Easier to Use, HBR
Opportunity: act as a bridge between data scientists and business people.
This course allowed me to dig further into the very interesting topic of Digital Twins, as I used the concept for my personal assignments.
What is a Digital Twin? It is a virtual copy of an industrial solution, an online replica of a physical system, used to track the past and predict the future. More than a static model, it is permanently enhanced with real-time data from the sensors of the live system.
Digital Twins are among the main use cases for the Industrial Internet of Thinks (IIoT), often referred to as Industry 4.0 or Smart Manufacturing. Their use is raising fast, and development environments are made available, originally from industrial suppliers (e.g. GE’s Predix), now more and more from software editors (e.g. AWS IoT Core, Microsoft Azure)
Wevolver has produced an insightful series of articles on the topic of Smart Manufacturing, covering the different IIoT components (data-driven production, networks and sensors, robots and cobots, …) and explaining with examples how IoT devices are to be used and deployed.
Opportunity: get started with Digital Twin technology by finding a suitable use case and playing around with IoT software platforms and development tools
A connected device alone is not sufficient to provide a successful IoT use case. It has to be connected to, and enhanced with, AI, Cloud, and Data. These 4 technologies together are the pillar of Digital Transformation, as Thomas Siebel very cleverly points out in his book Digital Transformation: Survive and Thrive in an Era of Mass Extinction.
“The confluence of elastic cloud computing, big data, AI, and IoT drives digital transformation. Companies that harness these technologies and transform into vibrant, dynamic digital enterprises will thrive. Those that do not will become irrelevant and cease to exist. If reality sounds harsh, that’s because it is.”
Thomas Siebel, Digital Transformation
Yet, as with any new technology, there exist barriers to a full scale implementation of innovative IoT applications, such as:
- Internal IT (runs the business, afraid from unknown things)
- Lack of clarity for cost-benefit analysis (no experience)
- Competition on talents
- Technological maturity
The potential for things going wrong is high. According to the IoT Industry Study by Cisco in 2017, 60% of companies underestimate the complexities of IoT and 75% of IoT projects fail.
But this is also why it is fascinating: going beyond the pure technical challenges and embracing a holistic understanding of the disruption at stake. And leading the necessary transformation initiatives in a professional, structured, yet adaptable approach.