Reading through analyst reports, one may be tempted to compare the hype generated by Industry 4.0 with the dotcom bubble we have experienced two decades ago. The internet revolution was a disruptive force that left an indelible mark on brick-and-mortar stores. Is it possible that the Smart Factory will have the same impact on industrial manufacturing? Will we experience a transfer of assets from industrial plants to IT solution providers?
The answer may be sought by looking at the fundamental differences in the roots of these phenomena. The Internet Boom was born in the euphoria of an optimistic California, whereas Industry 4.0 is firmly planted in the conservative German industrial heartland. In the late 1990’s, 27-year-old MBA’s became the prophets of the internet age with a message of new business models, company valuations and operational metrics. Driving today’s change are factory owners who measure yield rates and machine downtime.
Much of the change will be evolutionary and based on existing infrastructure. Change will be incremental and will not happen overnight.
According to a study by PwC, the industrial sector is expected to invest approximately 5% of revenue (US$904 billion) on the Smart Factory. The companies surveyed expect to achieve cost savings of 3.6% a year and annualized revenue gains of 2.9%. This translates into US$421 billion in cost reductions and US$493 billion in increased revenues per annum for a five-year period.
The Vendors’ ROI Pitch: Is it Too Good to be True?
Calculating a Return on Investment or ROI is as much of an art as it is a science, especially when estimates are based on scenarios that have not been tested. The easiest way to justify a large infrastructure investment is to forecast an even larger investment return. Underlying the move to Industry 4.0 is the additional revenue that can be generated by increasing industrial machine uptime. It is widely known that on average, 17 days of production are lost per year based on asset failure. The economic gain in additional productivity is significant enough to warrant investments in pricey solutions.
In the MIT Sloan Management Review, a GE executive justifies an investment in GE Predix with the following calculation: A single day of additional production at a Liquified Natural Gas (LNG) plant can generate $25 million and seven days of gained production can bring in an additional $150 million. With these types of potential returns, the investment in tens of millions of dollars in cloud infrastructure, new hardware and state-of-the-art sensors are relatively easy to rationalize.
Industry 4.0 is a Rational and Long Term Investment
The logic for investing in Industry 4.0 is concrete. An increase in worldwide petroleum production. A reduction in the Manufacturing Cycle Time. Industry 4.0 will likely have far-reaching implications, impacting the global economy, labor market and environment.
At the same time, there is resistance on the part of many industrial facilities to adapt nascent technologies.
Factory owners are overwhelmed with choice: leading technology vendors are announcing product releases at a dizzying rate without providing substantive proof for the efficacy of their solutions. It is difficult to make a large-scale and long-term commitment to an Industrial IoT platform that is based on relatively untested technologies. Adding a further layer of complexity is that fact that some of the leading vendors are unable or unwilling to share a definitive product roadmap, thereby forcing to factories to place a bet with limited information.
Apart from the issues relating to technology adoption, another constraint is the fact that many organizations lack the professional human resources to migrate to the Smart Factory. Although today we recognize the untapped potential of Big Data and Machine Learning, it is unclear how factories will be able to recruit the big data engineers and data scientists that are needed. For every new intelligent sensor to predict machine failure that is released to market, factory owners will need to find technicians to configure and monitor the system.
Don’t Shortcut the TCO Calculation
In recent years, Total Cost of Ownership or TCO has become a less popular metric for measuring IT projects. When selling the “big picture” TCO is often a nuisance calculation that misses the upside from an investment. Without negating the validity of this point, we should also consider some of the hidden costs of the Smart Factory.
There are numerous piecemeal solutions for predictive maintenance that have recently emerged. In each case, there is a requirement to purchase hardware, software and service fees.
On the other end of the spectrum, GE Predix, SAP Hana, and Siemens MindSphere are making massive R&D investments so that they can provide the underlying technology platform for the Smart Factory. We expect to see burgeoning partner ecosystems as more companies find ways to enter the lucrative IIoT market.
As the Smart Factory spurns innovation, TCO remains an effective tool to analyze solutions. Without ignoring the potential long-term impact, a thorough analysis of the total ownership cost is required. This includes setup costs, licensing costs and ongoing maintenance and support.
Presenso: Bridging the Gap between Cost and Value
Over the next couple of years, the tension between the long-term gain from Industry 4.0 and financially-driven decision making, will force factory owners to seek solutions that are cost effective.
Presenso upends the assumption that Machine Learning for Asset Maintenance requires the implementation of one of the large IoT infrastructure platforms that are gaining market visibility. In fact, with Presenso, the factory can delay the decision to adopt an IIoT platform because we offer Big Data and Machine Learning factory asset maintenance delivered as a service.
The Presenso Machine Learning solution is based on algorithms that require no information about a physical asset in order to detect abnormal sensor behavior and identify failure before it occurs. Second, Presenso is cloud-based and deployment occurs remotely. This reduces the need for onsite big data engineers and other skilled professionals. From a scale perspective, the Presenso solution continuously analyzes all the sensor data from an industrial plant in real time. Whereas some point solutions can monitor individual machines, with Presenso all the machine assets can be analyzed. Finally, Presenso does not require the purchase of new hardware such as Intelligent Sensors.
For factory owners that do not have the human capital or the financial flexibility to support the massive investment required for some of the emerging Industry 4.0 solutions, Presenso helps lower machine asset downtime without the risk of over-investment.
To learn more about Presenso big data solution for asset maintenance, please click here to schedule a demo.
Machine Learning Based Monitoring