From Failure to Success- A formula for Failing Forward

Author – Tim Smith

The first industrial revolution spanned 100 years, from 1750 to 1850. The second industrial revolution spanned 64 years from 1850 to 1914 ( wiki on industrial timeline ). The third industrial revolution ran from 1914 with the invention of the assembly line by Ford through to the introduction of automation. The fourth industrial revolution or Industry 4.0 is the application of digital technology to the shop floor. In the last twenty years we have seen a rapid adoption of digital technology driving IIoT, AI, machine learning, digital analytics all delivering on the promises of OEE, Lean, SixSigma, Throughput accounting and continuous improvement.

With the global market size predicted to reach USD 949.42 billion by 2025, Industrial IoT (IIoT) is expected to bring game-changing opportunities to improve operational visibility, efficiency, and productivity, all while lowering operational costs.

By 2030, Accenture forecasts that the IIoT could add $14.2 trillion to the global economy.

At least that’s the prediction.

A Cisco study revealed that only 26 percent of business decision-makers could name one of their IIoT initiatives a success. Beecham Research published a survey of IoT adopters which found that 58% of respondents stated their IoT project was either mostly unsuccessful or not successful.

According to a report by Bain & Company, difficulties in integrating IIoT solutions with existing operational technology and IT systems continue to be a major adoption barrier.

Digital innovation in manufacturing is so new that there is rarely an internal resource who has experience with similar projects. The top five challenges that deter manufacturers from even starting are complexity, lack of budget or staff, lack of knowledge, security concerns and indecision on what solutions will be most effective.

Gartner estimates 75% of IoT projects take up to twice as long as planned.  Customized, enterprise-wide initiatives that start at the top of the organization can take years to develop, configure and test before deploying into production, with prohibitive costs beyond the reach of most manufacturers.

Large internal IT projects run 45% over budget and deliver 56% less value than predicted. (McKinsey Digital). The skunk works approach rarely works in that, though a company may have IT and engineering talent, building a comprehensive, scalable, and extensible platform is never the main consideration in employing internal resources to accomplish an IIoT objective. In fact, what precipitates most failures is a chronically myopic approach to solving a problem or addressing a constraint. Such approaches are not the core competency of the staff of a manufacturer. Also, with multiple locations such talent may not reside at all of them, and the talent used to develop a skunk works solution can disappear taking with it all of the invested IP experience.

Industrial IoT promises exciting possibilities for factory automation, machine optimization and connectivity. But the onus is on the resources of the manufacturer. Many manufacturers simply aren’t equipped, and are left wondering why their IIoT aspirations fail, remain stuck in neutral or have never made it off the ground. The number of stalled Thingworx projects are vast as are large implementations of Siemens, GE, Dassault and others because of the massive disconnect between the solution builders and the target personnel. The lack of availability of development talent related to these proprietary systems leaves manufacturers scrambling to locate, attract, qualify, and hire such resources causing an HR nightmare. The only alternative is to engage the builder with uncapped paid services driving project costs through the roof. Additionally, the big-ticket vendors want to rip and replace everything under the guise of uniformity and simplified management, both of which rarely result from the process. A vendor should be considered a trusted advisor explaining how the United Nations of the shop floor, with a myriad of different machines, vintages and operations can be leveraged almost as-is in many cases and requiring only a single device on legacy equipment. In fact, when looking at a POC or even a facility install it should be quoted as a fixed cost project with a reasonably rapid implementation timeline with an ROI measured in weeks. This author has seen deployments of POCs in a couple of weeks and full plants in one to two months. Normally, the greatest amount of time expended in an implementation is having the manufacturer install network connections to the machine assets on the floor and setting up a shop floor network segment. (If you haven’t done this already, you better get busy!)

Many manufacturers decide to test the waters and deploy a proof of concept to better understand the challenges. The POC approach is supposed to alleviate large scale failure, however a POC approach is fraught with it’s own challenges. Without a clear business value and ROI, even a successfully deployed POC, still has no real quantification across the enterprise. The manufacturer who is running through a number of vendors looking to short list a few to then run a proof of concept is handicapped with lack of necessary technology and a lack of resources to commit to a successful POC. A faulty premise in pursuing a POC initiative is that there are too many platforms to test, or too many use cases to prove, or difficulty in justifying a business case without short term impact.

The author has talked to a number of manufacturers that must employ critical manufacturing engineers to babysit collection systems to keep them running or worst, accurate.

One of the biggest challenges to IIoT success is driving adoption and change management among users. The apprehension among employees to embrace new technology can even halt a project in its tracks. For Industry 4.0 to be successful, users must be highly engaged and collaborative on continuous improvement initiatives using new technology. Either a manufacturer must engage technologically astute shop personnel such as operators or select a system that abstracts the complexity away from the operators, providing simple interactive tools that do not threaten nor burden the operators with non-productive work.

Before implementing IIoT solutions, most manufacturers were manually collecting data on paper. While moving to automated data collection in an IIoT solution is efficient, it’s only as effective as the quality of the data that is gathered. There will always be a reason to enter some data manually. The problem is, many IoT technologies do not have a way to input data manually while also putting it into context. A system must be able to marry together the event data from the machines, the interaction by the operators and the flow of job information from the back office, without adding a layer of additional unproductive tasks.

The majority of IIoT projects are the brainchild of large technology vendors describing the utopia of grand worldwide connectivity. They sell the idea to upper management without due consideration of the complete lack of capability of the manufacturer to maintain or support it. IIoT project failures can be attributed to manufacturers not ready to process a nonstop firehose of high-velocity data with inputs from potentially millions to billions of data sets sent by connected devices located across the world. This data can and will arrive in a broad variety of formats and needs to be captured and processed in just milliseconds to provide valuable insights into factory and production conditions in real time. If one thinks about it, such abilities are beyond the scope of the majority of manufacturers. In fact, most multi-location manufacturers may not even be connected on a WAN backbone nor be running any type of collaborative, integrated services. Do not mistake the deployment and operation of an ERP system as a means of qualifying success when compared to the bi-directional, real-time movement of massive amounts of data across the enterprise. What a manufacturer needs is a system with inherent scalability to allow for planned managed growth, location by location.

You may find yourself facing any one of these challenges and that signals a stalled project. Whether that project is current or shelved, it’s time to take it up again and leverage your investment. You may ask yourself “why should I?” Digital transformation of the shop floor is commonplace and will become essential to stay competitive in a tightening Global Market. As companies realign objectives, re-shore plants and leverage resources in a shrinking pool of human capital, you will need to do everything in your power to stay ahead of the competition, or just keep up. You, or someone above you, decided that there is merit in connecting your shop floor and collecting data to drive improvement analytics. Unfortunately, the promise of your selected solution has not materialized for any number of reasons.

You need to step back and re-assess your current course of action by the following criteria:

  1. Fixed price project? If yes score 5 points, if no, score a 1
  2. Ability to connect multiple ways to machine assets? If yes score 5 points, if no, score a 1
  3. I have an internal champion to marshal resources? If yes score 5 points, if no, score a 1
  4. There is an out-of-the-box operator interface? If yes score 5 points, if no, score a 1
  5. Reporting and dashboards already exist? If yes score 5 points, if no, score a 1
  6. Risk assessment states deployment (per plant) in less than 3 months? If yes score 5 points, if no, score a 1
  7. There is a clearly defined ROI? If yes score 5 points, if no, score a 1
  8. The selected system can reside on-premise or in the cloud? If yes score 5 points, if no, score a 1
  9. IT group is on side and committed to the project? If yes score 5 points, if no, score a 1
  10. The shop floor personnel are committed to the project? If yes score 5 points, if no, score a 1

If you score 10 to 20, you need a serious re-assessment of the project and validate each of the above points and get them onside. You may want to review alternative solutions to bring the overall score up. If your project has overrun it’s expected deployment timeline or budget, then assess whether it was caused by external or internal reasons. If it’s external, then call a meeting and discuss what it will take to get the project finished. If item 2 is not a five, then you may need another solution. If it’s internal, then it usually means that items 3, 6, 9 and 10 are not fives and need your focus and attention to get them to five. If the issue is budget overrun, then assess whether there is a fixed timeline moving forward. If not, fire your vendor and seek a replacement vendor to finish the job.

If you score 21 to 30 you can probably restart the project by making sure that items 1, 3, 9 and 10 are fives. These items are critical for a reboot buy-in. Meet with your internal team, assign a champion, or replace the champion you have with someone who can liaison between all your departments including IT. Once you have your internal team recommitted, then meet with your vendor and get the job done.

If you score 31 to 50 then all you probably need is item 3 to be a five. If item 3 is a five, then it will probably be either item 1 or item 10 that has derailed you.

This article by no means is meant to be comprehensive or exhaustive. It is meant to get you thinking.

In short, the more of these assessment items are fives, the greater the potential of a stalled project finishing in success. What you do not want to do is leave the project on the shelf. Someone else will come along in a few months (not years), dust it off and take credit for it when it is successful.

Digital Transformation is already in your industry vertical. To survive, you must adapt.