Show me the data

Your Comprehensive Guide to IIoT Research

The Internet offers a plethora of information regarding IIoT initiatives; in practice, it can be difficult to separate the chaff from the wheat. In this article we present some of the most compelling findings about Maintenance 4.0 that can be used to make the business case for PdM in your plant.

For individuals tasked with developing an internal business case for investing in AI-driven Industrial Analytics, finding solid data to support this potential new expense can be difficult. Googling “what is the ROI for industrial analytics” reveals a hodgepodge of data and charts that aren’t necessarily helpful.  

To that end, we’ve compiled useful statistics that will help make the business case for Machine Learning for Industrial Analytics. Though we cannot vouch for the methodology used by each analyst, we have culled through numerous industry reports and identified research survey data to assemble the most comprehensive list available online. The data presented below should empower you to convince the different stakeholders of the value of a Machine Learning-based predictive maintenance solution. 

This article is divided into four sections that respectively address the following topics: 

Do you have the data to justify an investment in Industrial Analytics? 

  • The current state: what is the cause of industrial asset failure? 
  • Have you considered all cost factors in calculating the ROI of IIoT Predictive Maintenance? 
  • What are the tangible financial benefits of Industrial Analytics for Predictive Maintenance? 

Data required to understand the ROI of Industrial Analytics 

Let’s start with some of the more difficult data that may be indicative of the challenges your organization has. You may choose not to include these in the business case, but at the very least understand the challenges that lie ahead. At some point, you will be sitting in a meeting with an executive who will ask the following: “Before we make an investment, can you quantify the cost of asset downtime at our facility?”

Here are some basic statistics that you should consider: 

  • Downtime reduces an average of at least 5% of a factory’s productive capacity. In some cases, that number is 20%. 
  • More than 80% of manufacturers have no model to quantify Downtime Costs. Many industrial plants underestimate their downtime cost by 200-300%.  

The inability to accurately quantify downtime costs at your facilities is not an excuse to delay Industrial Analytics. In any Return on Investment (ROI) analysis of Industrial Analytics, an estimate for cost savings and revenue generated is necessary. 

There are two valid options: (1) Refer to industry data. In some cases, such as the automotive industry, it is has been estimated that downtime cost is $1.3 million per hour. In general, you are less likely to find updated information on pulp and paper, food and electronics manufacturing and more information on commodities such as oil and gas. 

(2) Model your assumptions such as revenue-per-facility and annual production costs. Even using a quick-and-dirty calculation is better than not answering this question at all. Your answer does not have to be correct to two decimal places, but your assumptions need to be valid and based on financial or operational data. 

Practitioner’s Guide: Cost Savings or Revenue Generation?  

When using statistics to build the case for Industrial Analytics, be careful about referencing the source of your data and explain the context. The fact that a research study was conducted on executives on the topic of Industry 4.0 or Industrial Analytics, does not mean that its applicable to your industry or industrial plant. An ROI calculation should be based on the projections for your specific environment. 

The Underlying Cause of Industrial Analytics Failure 

If Industrial Analytics for Predictive Maintenance is the solution, then we will need to address the underlying problem. Let’s take the oil and gas industry as an example. The two largest contributors to shutdown are Mechanical Breakdowns (46%) and Maintenance Related (23%). Out of the Maintenance Related shutdown, 92% were unplanned. 

In building the Business Case for a Big Data / Machine Learning solution for Asset Maintenance, it’s the 92% unplanned shutdowns that are the biggest opportunity.  

The Tangible Benefits of Industrial Analytics for Predictive Maintenance 

This is where your job gets tricky. If you have conducted a pilot for Industrial Analytics, then you are likely to have the data that will justify the investment. If you are not yet at the stage of the POC, then there is some information that may be helpful.  

There have been a few interesting studies that demonstrate the value of Predictive Maintenance programs. Keep in mind that Industrial Analytics is part of the Predictive Maintenance category. Without resorting to hyperbole, Industrial Analytics can be described as Predictive Maintenance on steroids.    

Therefore, the use of statistics that relates to Predictive Maintenance is acceptable as the data is likely to be conservative and even underestimate the true impact of adding Machine Learning to the Predictive Maintenance discipline.  

Here’s some useful data on Predictive Maintenance that was provided by the US Office of Energy Efficiency & Renewal Energy: 

  • Predictive maintenance program can yield savings of 8% to 12% relative to a program utilizing preventive maintenance alone.  
  • Cost savings for transition to Predictive Maintenance could “easily” recognize savings opportunities exceeding 30% to 40%.  

Surveys indicate the following benefits from implementing a functional predictive maintenance program:  

  • Return on investment: 10 times  
  • Reduction in maintenance costs: 25% to 30%  
  • Elimination of breakdowns: 70% to 75%  
  • Reduction in downtime: 35% to 45 
  • Increase in production: 20% to 25% 

If you are looking for data that relates to Industry 4.0, McKinsey has published a report with estimating the value of digitalization: 

  • 10%-40% reduction in maintenance costs  
  • Productivity increases of 3-5% 
  • 30%-50% reduction of total machine downtime  
  • CAPEX reduction of 3%-5%  

An Assessment of all Cost Factors Calculating the ROI of Industrial Analytics 

ROI calculations can be hard. As an engineer, Finance may not be your Mother Tongue Language. Focusing on key ROI elements (O&M and opportunity costs) could results in overlooking cost elements that are critical to the overall success of Industrial Analytics. 

The example below is based on oil prices of approximately $75 a barrel.   

We do not expect you to copy and paste this data, even if you are working in the petroleum industry. However, when you develop a business case keep this as a checklist for all the possible line items in your ROI calculation. Insurance premiums, productivity gains and capital expenditure reductions are line items that may be overlooked. Another ignored category (not specified in the table below), is the savings from lower injury rates. 

Here is where you need to proceed with caution. There are numerous studies that suggest that industrial plants are poised to make significant investments in Industrial Analytics for Predictive Maintenance. Perhaps some of these are accurate and reflect the future of Industry 4.0.

That said, basing investment recommendations on the buying behavior of industry peers is not advisable. First, what an executive may state in a research survey may not reflect the actions of his or her organization and there is no way to verify the veracity of survey data. Second, research reports are high-level and often lack context. Third, always check the source. Many research studies are US based or biased and may not be relevant to your market.

Nevertheless, we have included some survey research for you to consider in your Business Case. Caveat Emptor. Buyer Beware.

PWC Research Highlights

PWC surveyed 2,000 companies across the globe in 2016.  The following is an expectation of the impact of IIoT over the 5 years:

  • Average annual digital revenue increases of 2.9%. PwC calculated this to equal US$493 billion in increased annual revenues for the next five years across the industrial sectors we surveyed.
  • Average cost reductions of 3.6% per annum. In aggregate, survey respondents expect to save US$421 billion in costs each year for the next five years.

Honeywell Process Solutions (HPS) and KRC Research Highlights:

In the study sponsored by Honeywell Process Solutions titled, Data’s Big Impact on Manufacturing: A Study of Executive Opinions, the following findings were reported:

  • 67% of the 200 manufacturing executives interviewed will continue with plans in invest in data analytics over the next 12 months.
  • 42% indicated admitted that their equipment was being run harder than it should be.
  • 71% of respondents experienced occasional equipment breakdowns, while 64% reported occasional unscheduled downtime.
  • 40% ranked unscheduled downtime as the biggest threat to maximizing revenue.

IoT One Research Highlights:

This study by IOT One contains useful information about Industrial Analytics and is worthwhile reviewing:

  • The main application of Industrial Analytics in the coming 1-3 years are related to predictive and prescriptive maintenance of machines (79% of respondents view it as important).
  • Today, 68% of survey participants say they have a company-wide data analytics strategy, 46% have a dedicated organizational unit and only 30% have completed actual projects.



SKF Enlight AI

Industrial plants generate terabytes of process data. SKF Enlight AI is a SaaS Predictive Maintenance solution that uses Automated Machine Learning to identify emerging asset failure patterns within this data. It provides early warnings and sensor-level intelligence to help avert unplanned downtime and meet production goals. For more information on how SKF Enlight AI can improve performance and reliability, click here.

Leave a Reply

Your email address will not be published.