History shows that over 50 percent of those projects failed because the platforms were built without a solid understanding of what the user is going to do with them. Hence, start with a solid vision and business case of how IIoT will create competitive opportunity and profitability, says Michael Brooks.

 

The keys to getting started with IIoT should sound familiar: Be agile. Start small. Don’t wait, start now. Work fast on a real project. Learn, toss out what does not work and seek real business value. Okay, but what are the impediments to success?

 

IIoT projects can be today’s version of efforts made 20 years ago. Remember the IT group saying, “We need a data warehouse,” or “we need to be doing business intelligence?” My response remains the same today: “That’s the answer, what’s the question?”

 

History teaches us that over 50 percent of those projects failed because we built the platforms without a solid understanding of what we were going to do with them. You can see it happening again with IIoT, Industrie 4.0 and digitalization. So, don’t get lost in the zeal and enthusiasm for the latest technology; pick the most important business problem first and then work towards the solution. Do not build a special mousetrap then start looking for special mice. Start with a solid vision and business case of how IIoT will create competitive opportunity and profitability.

 

Culture and money

Your corporate culture decides if/how something can be achieved. Culture determines how your company will assess performance, allocate resources and propel intrinsic motivation. Is your leadership reaching past what was to achieve what can be? Has the funding been approved, not just for a trial, but for the real project? Is the organization ready? Most importantly, do your leaders show urgency – are they really walking the walk and totally aligned to improve operational excellence, safety, environment, downtime, profitability and more?

 

By selecting an important business issue in manufacturing or the supply chain you will do real work on a real problem. Clarity of vision and goals ensures you will avoid a science project where engineers and data scientists play technology games in the corporate sandbox looking for a problem. Drive a transformation from the executive suite implementing widely through P&L (profit-and-loss) business managers. Do not treat IIoT as an isolated IT implementation effort.

 

Avoid the competency gap

You can attempt to close the capability gap by building the skills and the culture to sustain it. However, my advice is to use products/tools that do the job and fit the competence and skills of current staff. The artisans on staff already understand the domain-specific issues needed for problem solving. Carefully assess your team’s experience and skills.

 

Avoid making impulsive hires who may be highly-skilled data scientists. Project implementers must know and understand use cases. Data scientists bring complex data analysis skills such as machine learning, but rarely deliver experience from the problem domain. Machine learning will find all manner of data correlations where some are often meaningless. For example, Figure 1 shows that importing Mexican lemons has improved highway safety a great deal.

 

Understanding causation requires knowledge and experience. What time, skills and experience will you need to attempt a solution, how long will it take, and will it scale? Choose carefully.

 

Technology, methodology and work process

You may discover that analytics technology alone is never enough to move the needle and drive business. Fundamentally, any IIoT project always needs technology, methodology and work process foundations. The path you choose to deliver these elements will determine whether your project is easy or hard - does the product you have selected embed methodology and work process for you? If not, your work will be far more difficult.

 

Make sure the technology is solid, tested, referenceable and not a science experiment. Then, if you have chosen wisely, your software application can also provide an implementation methodology that assures a solution is quick and easy to build with your team’s current skills and that it scales readily to meet site and corporate needs. Such competence will assure you can run a program to execute small and grow rapidly to scale – learning as you go.

 

If the application does not provide implementation guidance, then you must manually develop your own methodology. Similarly, the work process determines how you will use the finished application. What does it take? How do you respond? Is the process cumbersome, taking excessive time and effort? Can you handle it all yourself without donating intellectual property or needing lifelines for third-party experts and services? Is it important to minimise cyber security concerns by implementing the application completely inside your firewalls? You will benefit from the robust, proven application, its accompanying methodology and a work process that allows you to start small, execute and learn, and rapidly move and scale through your operations.

 

Data and software strategy

Avoid the trap of all dashboard, no improvement. Lagging indicators on dashboards cannot present the leading indicators of future performance that predictive and prescriptive analytics solutions provide. Avoid collecting ALL the data in case you might need it someday. Formulate your data requirements based on knowledge of the business and the problems you are trying to solve. We are talking about data-driven solutions. While you can always find opinions, no data means no solution. Make sure you understand what information you need for project success.

 

You are not doing banking, creating facial recognition, or building driverless cars… your correlations of symptoms and causes are in industrial data: messy, missing, highly volatile, and often mis-labelled. What data do you (and the application) need, from which sources, and at what frequency? How will the data be gathered, amalgamated, validated, cleansed? You must know these answers in advance to scope and scale the project. Last, is the solution truly data-driven, with high accuracy, always telling the truth with no false alarms? Is it capable of tuning itself to keep up-to-date as conditions change?

 

Partnering

Hopefully, you have come to realise that rather than a platform, the most important element in your IIoT project will be the application(s) that can assess the data, providing insight and foresight to drive the business. Choose the partner with the efficient solution that works for your business.

 

About the author:

Michael Brooks is the Senior Director, APM Consulting, at Aspen Technology. In a business development role, Brooks’ role is to accelerate the introduction of APM software suite to the marketplace.