Authors: Ani Bhalekar and Karel Eloot
A rapidly evolving vendor landscape composed mainly of point solutions makes it hard for industrial companies to build a scalable technology stack. Here’s how to cut through the complexity.
Nearly every global manufacturer seems to be testing Industrial Internet of Things (IIoT) technology, but in a recent McKinsey survey, nearly 70 percent of business executives said that their IoT initiatives are stuck in pilot purgatory, unable to reach company-wide scale. Furthermore, respondents said that only 15 percent of IIoT initiatives move to scale within one year, and a quarter of these initiatives extend longer than two years in pilot mode.
Clearly, this is not ideal, as pilots alone do not deliver the bottom-line impact companies need to remain competitive in the Industry 4.0 era. Even worse, because pilots do not encompass the full spectrum of what is possible with IoT-led technology transformation, these initiatives too often get cancelled because of the lack of a long-term vision and roadmap. But there is good news: designing the information technology/operational technology (IT/OT) architecture properly will enable companies to architect robust IIoT pilots that can rapidly scale.
Given that pilots often get bogged down because they’re too narrow to scale well, it’s imperative to design use cases so that they deliver end-to-end value for company-wide benefits. That also means addressing the overall IT/OT architecture design from the get-go to encompass the breadth of use cases. And with the IT/OT environment becoming more complex every day, the goal must be to sequence use cases based on two considerations:
- The priority of the use cases to the enterprise, based on predetermined KPIs that are linked to business value creation, and
- The ability of technology to support these use cases, while in parallel building both the technology stack (also known as the “platform”) and the applications specific to those use cases.
Industrial IoT stacks are developing in parallel with automation stacks
In our previous post, we described the factors that make constructing a scalable technology stack especially challenging in an industrial setting, such as the complexity of the industrial-automation ecosystem, the alphabet soup of sensor subsystems, legacy machines that are still unconnected, and poor collaboration between IT and OT departments.
Consequently, industrial companies often struggle to determine how best to source a solution that truly unlocks value from data and analytics. Vendors of all types are introducing new technology to capitalise on this trend, as they seek to secure a portion of the multi-billion-dollar market (exhibit).
At one end, traditional industry-automation vendors that are established incumbents, with strong positions in control points, are introducing complementary IoT solutions to help manufacturers deploy an end-to-end IIoT platform. The playing field also comprises established vendors from both the OT and IT stack, each with strengths in different technologies. Some vendors are attempting to create portfolios that span most of the IT/OT automation stack, while others are steadfast in their commitment to a focused area of expertise.
We looked at five of the top automation vendors and found that most are expanding their IIoT portfolio and capabilities. The overall landscape thus remains unsettled. For example, just among this vendor subset, there are almost 50 proprietary offerings in more mature areas of the stack, such as manufacturing execution systems and similar technologies—and standardisation across industry use cases has yet to take off. We also found that several major vendors lack offerings in areas such as building automation, human-machine interfaces, and sensors and RFID.
Traditional software vendors are also trying to solve IIoT issues
To add another layer of complexity, traditional software vendors are competing for their share of the industrial-cloud ecosystem. Most approach the ecosystem from bottom up, with initial success as an infrastructure-as-a-service (IaaS) provider, leading to vertical integration in platform-as-a-service (PaaS), and finally leading to non-industrial software-as-a-service (SaaS). The natural extension is then to enable industrial SaaS.
As with the automation stack, there are several major players competing but there is no clear leader. For plant operators, it’s therefore nearly impossible to predict who will come out on top, leaving internal teams to navigate a complex landscape of evolving technology offerings from vendors with whom they have an established and amicable relationship—both on the IT and OT side. The only clear message right now is that no single vendor will likely be able to solve all of a company’s requirements.
A proven approach to developing a scalable IIoT stack
Knowing your starting point and destination can help make even the most complex journey navigable. This is true of developing an IIoT stack capable of scaling. We recommend our clients take the following steps when initiating an IIoT pilot. Over the next few blog posts, we will deep-dive on each of the following steps with examples where applicable.
- Start by generating a solid list of use cases. Focus not only within the four walls of a given plant, but also extend into supply chain and design, with a view to enabling digitisation of complete value chains. This exercise doesn’t have to be exhaustive, but it should be representative of the things the stack needs to be able to do for you. For each use case, you can then determine the tech-stack requirements, which collectively will show what the tech stack will have to deliver to achieve scale at the right speed and cost.
- Develop a future-state reference architecture based on business need. Begin with an assessment of your current IT/OT stack, from plant level to the whole enterprise, looking at everything from technology architecture to manufacturing applications and tools. Also review previous IIoT pilots, identifying pain points that prevented projects from moving forward.
- Incorporate data expertise up front. Avoid delays and needless iteration during implementation by having data scientists, who know the data they will need to have when building their analytics models, participate in the design of the reference architecture. The data scientists will also need to work with the process engineers to ensure that the data can be both collected accurately at source and aggregated into the right location at the right time.
- Define a core architecture choice. Assess the pros and cons of selecting a single platform, an ecosystem of vendors, or a hybrid approach, before aligning on an overall direction.
- Recommend tech choices specific to use cases. Agree on tech-stack requirements for prioritised use cases and compile decision criteria. Tailor existing evaluations of IIoT platform providers against the decision criteria to identify top choices.
- Chose vendors based on both tech capacity and human capability. While getting the right tech is essential, getting it to work depends on system-integration capabilities as well.
- Deploy technology for prioritised use cases in the short term. For the prioritised use cases, invite a curated list of vendors into your organisation’s ecosystem.
- For the long term, create and align IT/OT roadmaps. Draft consistent IT and OT roadmaps that align with your organisation’s broader business goals and underlying tech, then syndicate and refine them—noting that the time scale will be very different from a multi-year ERP implementation.
Following this approach will help your company navigate the complexities of the IIoT landscape to capture value quickly. Watch this space for more on each of the steps listed above.
The authors would like to thank Mike Coxon, SubuNaranayan and Bodo Koerber for their contributions to this post.
About the authors:
Based in McKinsey’s Singapore office, Associate Partner Ani Bhalekar is Vice President for the Internet of Things (IoT), serving clients across sectors on smart cities, connected transport, Industry 4.0, analytics, mobile and digital technologies.
Focuses on broadly improving operations in industrial sectors, particularly for companies in China.