There isn’t a week without the state of global manufacturing being discussed and scrutinized in media headlines.
In the past, productivity gains stemmed from increasingly standardized processes. This led to a global industry characterized by mass production, in an effort to achieve economies of scale, geographic fragmentation, to source the cheapest resources and complex supply chains, which paved the way for third and fourth-party logistics service providers.
As the sector continues to evolve, to what extent will digital manufacturing become the next driver of large productivity gains? After digging into the space, these are some of our key takeaways and market theses.
1. Some verticals will become more automated, while others will become less.
For most, this seems highly counterintuitive. Headlines constantly point to the same conclusion: robots will replace humans entirely. The problem with this statement is it assumes all manufacturers have the same production methods and thus, have the same pain points. Research shows otherwise.
To uncover underlying drivers for automation, we need to understand the three main types of manufacturing methods: repetitive, batch and job production. As the name suggests, repetitive production involves heavily automated manufacturing where the goods are identical. An example of a repetitively produced good would be frozen pizzas or silk fabric. In contrast, job production is done on a good-to-good basis, where items are crafted individually at work stations. Typically, there is little to no automation, but a high level of customization, as seen in apparel and aerospace factories. Batch production falls somewhere in between; processes start off fairly repetitive, but at some point the assembly lines diverge to create slightly different variations of the same product, as seen with automotive and furniture production.
To say that a car manufacturer and cosmetics producer seek the same level of automation is overly simplistic. Today, and for the foreseeable future, humans play a vital role in production. When it comes to process flexibility, humans trump machines. Not only is being adaptable to task variability an asset in crafting highly customized products, but it’s critical for when changes to the assembly line are made. This means that human involvement is important for batch production and critical for job production.
Because of this, some verticals are becoming less automated than they used to be. In 2016, Mercedes-Benz started swapping out its assembly line machines for more skilled workers. With product customization playing an increasingly important role in gaining market share, the second-largest luxury car manufacturer turned to “robot farming” instead. This involves equipping humans with hand-held robots that can automate some tasks. Markus Schaefer, Mercedes-Benz’ head of production told Bloomberg: “Robots can’t deal with the degree of individualisation and the many variants that we have today. We’re saving money and safeguarding our future by employing more people.” Toyota is following suit. Across verticals, and depending on production processes, we’ll continue to see manufacturers make the tradeoff between automation and production line flexibility. Leading automation providers, such as Clearpath Robotics, are designing advanced automation systems to augment human workers without compromising on flexibility.
2. Industry-wide adoption of additive manufacturing and VR/AR isn’t imminent… Yet.
Additive manufacturing, VR and AR have been prominently featured technologies in headlines for some time now. Countless articles claim they are “game changers” for various industries, including manufacturing. But, innovation and disruption are not one in the same. In order for a technology to be truly disruptive, it needs to achieve mass adoption. As it stands, all three technologies face important roadblocks to become mainstream.
Additive manufacturing, also known as 3D printing, has made notable progress, but is still years away from debuting in the average factory beyond prototyping. While the level of printing accuracy has increased substantially and the list of printable materials keeps growing, concerns surrounding the cost-to-quality ratio remain common. Industrial level printers cost anywhere from $250-$850k, while operating them requires 50–100x more electrical energy than alternative production methods used today. Not the most economical investment. From a quality perspective, many BPA-filaments currently aren’t printable. One example showed 3D-printed utensils containing empty spaces for bacteria to easily grow, raising questions as to whether it meets health and safety regulations. Continuous improvements from open source communities and patent expirations (e.g. a 2009 expiry dropped the price of fused deposition modelling printers from $10k to $1k) will fuel the commoditization of 3D-printing hardware in the coming years while true differentiation will happen at the software layer.
VR/AR have their own set of barriers. As it stands, their most common use cases pertain to R&D, labour safety and data collection, such as fitting workers with glasses that scan barcodes. But, a PwC survey of 120 major manufacturers found that almost two-thirds have either not yet adopted or have no plans of adopting either of the technologies. The reason? Roughly 30% say it’s not the “prime time” for adoption, while others reference cost, weak use cases, complicated deployment and lack of benefits as obstacles. The killer app for VR/AR in manufacturing is yet to be defined. Until these technologies can fully justify their costs of acquisition and implementation to provide a clear ROI, it’s hard to envision a breakthrough to mainstream adoption.
3. Industrial IoT is the most lucrative opportunity in digital manufacturing, but it’s more complicated than you think.
Industrial IoT (IIoT) remains promising for both established and emerging players tackling the space. Machine-to-machine (M2M) communication is pushing the next frontier of automation, where factories could, theoretically, operate entirely independently from humans. For dangerous industries, such as energy and mining, this is a huge selling point. For others, it’s the predictive analytics, real time KPIs and remote management use cases that are most appealing.
IIoT has been gaining attention from a growing ecosystem of potential customers, investors and acquirers. The manufacturing sector already uses IoT applications the most, accounting for approximately 35 percent of total market users in 2016. By 2025, the total market value is expected to balloon to ~$930B, up from ~$130B today. This is largely attributed to the hefty investments manufacturers are currently making. According to Accenture and GE, almost 75 percent of global companies with revenues greater than $150 million are investing over 20 percent of their tech budgets in IIoT. When asked whether their investments will increase over the next year, again, around 75 percent said it would.
But, machine connectivity isn’t simple. It’s far from it. To enable things like automatic KPIs, a complex infrastructure must first be put in place. Assuming the average plant already operates with basic hardware, such as machinery and sensors, the journey to connectivity starts with the network. Industrial wifi can be a huge upfront investment, with deployment costs averaging $250k. On top of that, network management and cybersecurity features are often needed to ensure smooth operation. The next layer enables edge processing via edge device gateways. The commonly-used industry buzzword refers to the ability for data to be processed at the source. This, as opposed to aggregating data from multiple sources, paves the way for a more intricate early-stage filtering procedure. Since an estimated 90 percent of data lakes are ultimately useless, being able to capture just the important information is a huge value-add. This data is then stored in the cloud, where it can readily be accessed. The typical cloud-based analytics provider can then track KPIs in visual dashboards or generate predictive analytics. Some providers will even offer SDKs to deploy your own applications.
The takeaway here is that getting production insights is hard. It requires substantial investments in the underlying architecture that won’t be made overnight. The most successful players in this space will put an emphasis on easing the transition for its customers, either through flexible value propositions or supportive partnerships for end-to-end solutions.
To sum up, navigating the digital manufacturing landscape can be a daunting task. But, hopefully these predictions and myth-busters will help you interpret readings more easily and evaluate digital manufacturing startups opportunities with a healthy dose of skepticism. In a market all about timing, it appears this could be the next big wave. And if you’re patient and watch closely, you may be able to catch a ride.
This article was originally published on iNovia Capital’s blog. Copyright 2017.
