Razor Insights

Mythbusting Connected Factories in 2024

Debunking Eight Common Misconceptions in Digital Manufacturing

Unlock the truth about connected factories in 2024! We're debunking the top eight myths in the industry and revealing the essential nature of digital manufacturing.

<>

We’re in the 25th year of this century. The world population is 8.119 billion; getting payloads into Earth's orbit is the cheapest it's ever been; Big Data is a global market worth around $100 billion and PwC expects AI to add $15.7 trillion to the global economy over the next 5 years. Meanwhile…incumbent industry players are digitising at a rate never seen before.

Here, Razor has cut through some myths around factory connectivity and is sharing its hard-earned expert knowledge to help those considering dipping their toe into the world of digital manufacturing.

Ready? Let's get dipping!

<>

Myth 1: “A connected factory is a ‘nice to have’, rather than essential.”

The reality is that both domestic and international competition are eroding margins for laggard UK manufacturers. Even for high-value products that involve highly capital-intensive processes and highly skilled, expensive labour - many countries are rapidly catching up. 

It is no longer sufficient to have a healthy price margin on products, but also essential to maximise the utilisation of your assets, significantly improve quality (enhancing customer satisfaction as well as enjoying financial benefits of greater yields), and endeavour to maintain accurate data (which can now include terms such as logistical overheads, energy use or carbon emissions).

It’s widely known that the UK has a productivity problem, and the route out of it has always been a culture and business structure that embraces continuous improvement. Here, digital technologies are principal enablers - a connected factory directly drives capability for continuous improvement and can begin to replicate advantages found in greater automation - higher productivity, process standardisation and greater asset utilisation. 

There is also a soft benefit - your manufacturing company is seen as modern, supporting training, retention and acquisition of great employees for now and the future. Employees are released from mundane tasks to work on more interesting, more rewarding, higher-skill, higher-value problem-solving, armed with data and skills they need to improve operations.

<>

Myth 2: “Getting a connected factory costs a fortune in time and money”

Business investments must be justified, risk needs to be managed and the return on investment (ROI) needs to be clear. Fortunately, approaches for digital transformation need not be exclusively “big bang” - many, small, focused investments are more likely to succeed - something Razor has advocated for many years.

An example of this in connected factory transformation is to select specific assets or processes that are seen as critical - those that are highly utilised (must be running continuously, perhaps justifying the use of predictive maintenance technology), those that are undertaking finishing or inspection operations, (where accuracy is paramount, labour costs high and manual data collection is sporadic or challenging) and finally - those assets that are not performing well - here information and data can offer insights that will allow steady, incremental improvements to this asset to bring it in line with the rest of your facility. 

Many software offerings are OpEx, incurring a monthly or annual subscription fee which avoids managing a CapEx investment and encourages a cloud-based Software-As-A-Service (SaaS) implementation, with continuous development of new features. In some cases, this also allows for general access to the platform over the internet from any device and from anywhere. The majority today offer a hybrid of the two - Razor’s own MIP framework is no exception - offering the best of both worlds.

many small focused investments are more likely to succeed

<>

Myth 3: “The advantages of asset connectivity are not justified by the integration overhead”

It’s true that pure software implementations - such as paperless systems, avoiding manual entry of data or providing digital work instructions are often well-priced and easily deployed, avoiding the overhead of integration in cyber-physical systems. That said, older assets can be retrofitted with proprietary sensors, and awareness of something as simple as “utilisation over time” can enlighten manufacturing leaders to where bottlenecks within the manufacturing system lie, or where greater capacity can be unlocked. Labour costs are some of the highest - and are also easily calculated - so any projects that can reduce labour requirements are often easily justified.

Newer assets may offer up a plethora of data points that can characterise the manufacturing processes, supporting higher quality levels and even automating inspection - but only in the most stable manufacturing systems are these systems likely to demonstrate high ROI.

connected factories through the industrial internet of things provide access to data on an unprecedented scale

<>

Myth 4: “The more data collected, the better.”

We acknowledge that collecting data is key and a huge enabler of digital transformation. Connected factories, through the Industrial Internet of Things (IIoT), provide access to data on an unprecedented scale. 

The common misconception is that a wide range of data points - almost anything - at the highest volume possible is the basic first step. In a nutshell - “collecting data” is sufficient. 

On the face of it, this ‘bottom-up’ approach is correct, in that, without data, no claims made within this myth-busting session have any real applicability. However, deciding what is “good data” to collect deserves consideration and planning. Selection of the data must come from business value - what data points are important – which can derive actionable insights? In a sense, we’re saying that you need to work backwards from possible insights to evidence!

Software is more than capable of processing huge volumes of data, but the use cases need to be clear. An example of this could be anomaly detection for predictive maintenance or the deployment of computer vision - here, masses of data are processed “on-edge”, and ‘small data’ is output - perhaps the estimated health or a geometry estimation, respectively.

Razor has led thinking around the effective use of data in businesses generally, find out more here.

<>

Myth 5: “The data available from assets is only utilisation”

Make no mistake - even establishing utilisation data from assets would be a huge step forward for many manufacturers. The revelation that many of their assets have fractional utilisation means that manufacturers can step up their planning and scheduling ability, potentially reduce product/part costs, increase production volume and grow their market share.

In most cases, what limits data from assets is the age and digital maturity of said asset. Whilst many older CNC machine tools will offer data from their NC Controller, others will need to be retrofitted with sensors.

Razor pioneered the concept of using computer vision to simply ‘observe’ the asset in real-time - and for assets that are larger or more niche, this can be a hugely effective connected factory play. Again, this simple data would be machine state; “running” or “stopped”, but in more advanced applications, visual data can offer a way of characterising the process itself or estimating part geometries whilst in-process.

older assets can be retrofitted with new sensors to gain awareness of metrics such as utilisation over time

<>

Myth 6: “Smart factories are insecure”

Connected factories are loosely related to Internet of Things (IoT), and in the IoT space there are numerous accounts of security vulnerabilities where poor engineering practices as well as poor system design that can be exploited. 

However, since the inception of IoT,  the general growth of Cloud and the ubiquity of IT systems, many, many enterprises, security vendors, startups, and device manufacturers have made IIoT security a priority. Furthermore, security standards bodies are aggressively developing processes and security frameworks for secure identification of IIoT

Nevertheless, the decision to use an on-premise deployment or a cloud deployment of a connected factory will depend on a variety of factors, with security trade-offs being a large part of this decision. Some engineering companies will need to retain specific data on-premise - the use of cloud is out of the question.

<>

Myth 7: “Industrial monitoring systems are only for large businesses.”

One common misconception is that Industry 4.0 applies only to large corporations, with a prerequisite of deep pockets to invest in new, state-of-the-art systems. A sweep of the industry publications is littered with examples of blue chip manufacturers investing huge sums of money into new facilities and digital technologies. This can give the impression that this world is exclusive to these big players. 

This is not the case.

Whilst larger companies have the benefit of greater resources, better prepared in terms of digital literacy and often experienced leadership, small and medium-sized enterprises (SME’s) have the advantage of lean decision making, a flatter organisational structure, a smaller communication overhead and less bureaucracy. All traits sought by digitally transforming businesses. 

Even with constrained resources, implementing a digital strategy doesn’t have to mean ripping out “what works” (existing systems) with complex and expensive infrastructure that claims major steps forward. Razor has a long history of supporting businesses of many types and understanding what digital transformation means to them - a roadmap for the enterprise that aligns with the long-term business objectives is essential. Read more about how the principles of Enterprise Architecture (EA) work to achieve this here.

<>

Myth 8: “Industrial monitoring systems are there to establish a dashboard.”

Arising from experience in establishing all the required facets in new premises, it is believed that any path towards implementation of an informatics platform is far easier when you're working on a clean slate. 

This is at least partially true in the case of IT infrastructure, equipment and assets. Older sites are often not networked, and older machinery does not offer the type of connectivity required for natively sharing data. 

Industrial monitoring systems cannot however be equated with fully automated, lights-out manufacturing systems. Brownfield sites can be systematically improved with industrial monitoring systems - the older, unconnected assets can be fitted out with proprietary sensors. The decision here depends on how mature the implementation needs to be - is it only asset utilisation that is of interest? Or is it high-fidelity process monitoring?

A lack of appropriate infrastructure can in some cases be addressed by implementing a new network - including using WLAN and in some cases private 5G.

<>

In Conclusion

Connected factories are critical to high performance, and it is not the largest or the most modern manufacturing systems that are exclusively capable of digitising. Nor are they some mystifying, dark art - as of 2024, clarity of why, what and how to get started is clear.

Start with a pilot project, understand what measurement means for your business and create a plan around what data provides this information. Prove value and iteratively increase your scope. 

For advice on making it happen, and where to start, Razor and our extensive network of contacts can help. Get in touch to find out more.

Diving deeper into Razor

What we think