Companies are generating ever increasing amounts of data associated with business operations, leading to renewed interest in predictive analytics, a field that analyzes large data sets to identify patterns, predict outcomes, and guide decision-making. Companies are also facing a complex and ever expanding array of operational risks to proactively identify and mitigate. While many companies have begun using predictive analytics to identify marketing/sales opportunities, similar strategies are less common in risk management, including safety.

Classification algorithms, one general class of predictive analytics, could be particularly beneficial to the refining and petrochemical industries by predicting the time frame and location of safety incidents based on safety related inspection and maintenance data, essentially leading indicators. There are two main challenges associated with this method: (1) ensuring that leading indicators being measured are actually predictive of incidents, and (2) measuring the leading indicators frequently enough to have predictive value.

Kestrel’s article in the Q3 2018 edition of Petroleum Technology Quarterly (PTQ) features a case study to illustrate this process. Using regularly updated inspection data, the author developed a model to predict where broken rails are likely to occur in the railroad industry. The model was created using a logistic regression modified by Firth’s penalized likelihood method, and predicts broken rail probabilities for each mile of track. Probabilities are updated as additional data are collected.

In addition to predicted broken rail probabilities, the model identifies the variables with the most predictive validity (those that significantly contribute to broken rails). Using the model results, the railroad was able to identify exactly where to focus maintenance, inspection, and capital improvement resources and what factors to address during these activities. Validation tests of the model revealed 70% of the actual broken rail incidents occurred on the 20% of segments at highest risk for broken rails.
The same methodology could be used in the refining and petrochemical industries to manage risks by predicting and preventing incidents, provided that organizations:

• Identify leading indicators with predictive validity
• Regularly measure leading indicators (inspection, maintenance, and equipment data)
• Create a predictive model based on measured indicators
• Update the model as data are gathered
• Use the outputs to prioritize maintenance, inspections, and capital improvement projects and review operational processes/practices.

Read the complete article in PTQ.

Traditional processes tend to produce traditional results. You can’t expect technological innovation without technological integration. The key is identifying the traditional processes that yield benefits (most likely cost or time savings) from technological integration. Doing this allows companies to stretch and empower every limited resource.

Fix It & Find It

Take the business practice of internal auditing as an example. The most traditional practice for internal auditing (e.g., environmental, safety, DOT compliance, ISO 9001, food safety) is a “find it & fix it” cycle, where the internal auditor goes out into a facility and audits operations as they exist. During the audit, the auditor identifies issues based on a standard set of protocols. The auditor typically walks a facility with a notepad and pencil taking notes of field observations that aren’t in compliance. Following the audit, the auditor creates a report and shares the findings with a responsible party. This can take weeks or even months. The cycle is repeated when the auditor comes back at a later time to check the site again.

The “find & fix” audit cycle works, but only to a point. The difficult part comes next. What happens with that inspection form or accident investigation report after it is completed? It is likely reviewed by a few people, perhaps transcribed into electronic form by a data entry clerk, and filed away someplace for legal and compliance reasons, rarely (if ever) to be seen again.

Filing data away in a drawer is better than nothing because it does show some documentation of findings, but that is where the benefits end. What happens when the auditor is asked to compile year-long data from the findings? How do you evaluate patterns and trends to best allocate your limited resources for improvement initiatives? The paper method of recordkeeping makes compiling field data into a report an enormous task.

Electronic Data Capture

If the auditor were to capture all the field data via smart phone or tablet at the point of discovery, the task of generating a report to analyze trends would be much easier. When data is collected, uploaded, and stored in a database, accessing and reporting on the data becomes as easy as asking a question. Questions like, “How many deficient issues were there at the warehouse last year?” or “How many overdue action items does Bruce have in repackaging?” can be answered by simply making a request of your data.

Data entered in the field can be used in many ways. Some applications written for devices allow you to print reports immediately from the smart phone and tablet device. Others require the data to first be uploaded to a desktop computer. Either way, the reports generated can include photos at the point of discovery and reference information, along with field comments. These reports support the auditor’s findings and remove questions about what was observed or whether a situation is in violation of the protocol. Subsequently, these reports also become a valuable learning tool for employees in the field.

Once uploaded, the data is stored in a database for later reference. Assessments continue to be added as audits are performed to amass a large bank of data. In electronic format, that data (unlike handwritten notes) can be easily arranged for analysis. Reports can be generated using a large menu of criteria, including running statistics on a site over a period time or identifying instances of a certain violation. Mining your data in different ways helps identify root causes and end harmful trends so that real improvement can occur.

This is the third in a series of articles on food product recalls.

One form of protection from the economic and reputational damages of a food recall event can be to transfer risk through an insurance policy that is specifically designed to respond to a recall event.

Are You Covered?

Business Owner’s Insurance Policy “BOP” provides most enterprises with two main forms of coverage: Commercial General Liability (CGL) and Business Property. Many food and beverage companies believe that basic CGL insurance coverage will provide protection in the event of a product recall. In reality, CGL policies typically contain an exclusion (Recall of Products, Work or Impaired Property) that precludes coverage for any claims associated with a product recall or withdrawal.

Because most CGL policies do not cover recall-related losses, separate Product Recall, Business Interruption, or similar types of insurance can provide protection to reduce the potential financial impacts of a recall event. Companies can purchase either first-party or third-party Product Recall policies, or both.

First-party policies provide coverage for the company’s own economic loss incurred due to a recall. These losses may include:

• Business interruption
• Lost profit
• Recall expenses
• Expenses to respond to adverse publicity and rebuild a brand’s image
• Consultant and adviser costs

Third-party coverage applies to economic loss incurred by third parties (e.g., distributors, wholesalers, customers) who may be impacted by a recall event. This could include broad coverage for numerous costs associated with the following:

• Removal of recalled product from stores
• Transportation and disposal of the product
• Notification to third parties of the recall
• Additional personnel/overtime
• Cleaning equipment
• Laboratory analysis

Business Interruption insurance is another coverage that may cover not only catastrophic losses, but also food recall events. If purchased, it is important to make sure that the Business Interruption coverage works hand-in-hand with Product Recall coverage.

What to Look for in a Policy

Product Recall insurance should be specifically tailored to meet the needs of the company. Here are some things that a company should ask when exploring Product Recall/Contamination insurance:

• Will the policy cover recalls where there is limited likelihood of bodily injury (e.g., class II or class III recall that is less severe)? What if a recall is requested (vs. ordered) by the FDA or USDA?
• Will the policy cover loss from an FDA administrative detention?
• What happens if the company experiences financial loss due to a recall and then the facts underlying the recall turn out to be incorrect? Are those losses still covered?
• Does the policy exclude coverage if the recall was due to a problem with a competitor’s product? What if the product breaches a warranty of fitness?
• Does the policy provide coverage for claims by third parties (e.g., customers)?
• Does the policy cover lost profits/revenue? What about logistics and repair costs (e.g., shipping and destruction, public relations, product replacement, and reputation/brand damage)? How is the loss calculated?

*****
Read:

• Part 1
• Part 2

Chemicals are an important part of many aspects of our lives; however, improper handling and management of chemicals can result in catastrophic releases that have severe and lasting impacts—loss of life, injury, property damage, community disruption.

The USEPA’s Risk Management Plan (RMP) Rule (Section 112(r) of the Clean Air Act Amendments) is aimed at reducing the frequency and severity of accidental chemical releases. While the intent of the RMP Rule is positive, there has been much controversy over what the rule requires. This has resulted most recently in the RMP Reconsideration Proposed Rule, which was published on May 30, 2018.

The History of Modernizing RMP

RMP regulations were first created in 1996 to protect first responders and communities adjacent to facilities with chemical substances. Changes to the original RMP Rule have been in progress since former President Obama issued Executive Order (EO) 1365, Improving Chemical Safety and Security, in August 2013. Modernizing policies and regulations—including the RMP Rule—falls under this umbrella.

A July 2014 Request for Information (RFI) sought initial comment on potential revisions to RMP under the EO. This was followed by a Small Business Advocacy Review (SBAR) Panel discussion in November 2015. On March 14, 2016, the USEPA published Proposed Rule: Accidental Release Prevention Requirements: Risk Management Programs Under the Clean Air Act, Section 112(r)(7), outlining proposed amendments to the RMP Rule.

The much anticipated final RMP Amendments were published in the Federal Register on January 13, 2017. According to the USEPA, these amendments were intended to:

• Prevent catastrophic accidents by improving accident prevention program requirements
• Enhance emergency preparedness to ensure coordination between facilities and local communities
• Improve information access to help the public understand the risks at RMP facilities
• Improve third-party audits at RMP facilities

After the USEPA published the final rule, many industry groups and several states filed challenges and petitions, arguing that the rule was overly burdensome, created potential security risks, and did not properly coordinate with OSHA’s Process Safety Management (PSM) standard. Under the Trump administration, the USEPA delayed the effective date of the rule until February 2019 and announced its plan to reconsider the rule’s provisions.

Reconsideration

That brings us full circle to the RMP Reconsideration Proposed Rule that was published at the end of May. According to the USEPA, this reconsideration proposes to:

• Maintain consistency of RMP accident prevention requirements with the OSHA PSM standard.
• Address security concerns.
• Reduce unnecessary regulations and regulatory costs.
• Revise compliance dates to provide necessary time for program changes

What’s Going?

USEPA Administrator Scott Pruitt said in a press release, “The rule proposes to reduce unnecessary regulatory burdens, address the concerns of stakeholders and emergency responders on the ground, and save Americans roughly $88 million a year.”

To accomplish this, the reconsideration proposes making the following changes:

• All accident prevention program provisions have been rescinded in the reconsideration so the USEPA can coordinate revisions with OSHA and keep regulatory costs in check. This includes repealing the requirements for conducting:
  • Third-party audits
  • Safer Technology and Alternatives Analysis (STAAs) as part of the process hazard analyses
  • Root cause analyses as part of an accident investigation of a catastrophic release or near-miss
• Most of the public information availability provisions have been rescinded due to their redundancy and security concerns, particularly regarding specific chemical hazard information. The USEPA is proposing to retain the requirement for facilities to hold a public meeting within 90 days of a reportable incident.

What’s Staying?

Many of the emergency coordination and exercise provisions of the Amendments rule are staying–but are being modified to address security concerns and provide more flexibility. The Reconsideration Proposed Rule still requires facilities to:

• Coordinate response needs at least annual with local emergency planning councils (LEPCs) and response organizations, and to document these activities
• Provide emergency action plans, response plans, updated emergency contact information, and other information necessary for developing and implementing the local emergency response plan to LEPCs
• Perform annual exercises to test emergency response notification mechanisms (Program 2 and 3 facilities)

Looking Ahead

The proposed rule is available for public comment for 60 days after its publication date (May 30, 2018). In addition, a public hearing is scheduled for June 14, 2018. If the Reconsideration Proposed Rule is published, compliance dates will be as follows based on the effective date of the final rule.

For more information, visit the USEPA website on the RMP Reconsideration Proposed Rule.

This is the second in a series of articles on food product recalls.

The risk for all food companies of being affected by a recall is substantial—and to adequately respond to a recall claim is complex and expensive. Companies should always be prepared to prevent a recall from occurring.

Here are seven tips that can help your company prevent and/or manage a food recall:

1. Establish Food Safety Plan using HACCP approach or preventive controls. Always make sure the plan is kept up to date with facility production or product formulation changes to ensure potential risks are controlled.
2. Develop and maintain a written Recall Plan, as well as a Crisis Management Plan. These plans should be reviewed, tested, and updated at least annually. Lessons learned should be recorded and analyzed for possible improvements.
3. Conduct mock traceability exercises over a certain time period. In case anything occurs from within the supply chain, you should be in control of your own ingredients and finished products.
4. Establish a functioning approved supplier program.
5. Utilize third-party audit certification to establish a Food Safety Management System (FSMS), and gain senior management commitment and resources for maintaining the FSMS onsite. This may be in the form of commitment to a Global Food Safety Initiative (GFSI) benchmarked standard (e.g., BRC, SQF, FSSC 22000, IFS).
6. Implement thorough sanitation and hygiene processes.
7. Maintain all related documentation and records.

Read:

• Part 1
• Part 3

The new ISO 9001:2015 standard for Quality Management Systems (QMS) was issued in late 2015—which means the three-year transition period to become certified to the new version is now in full swing. Change can certainly present challenges; however, the ISO 9001:2015 update is designed to simplify the requirements, focus more on business needs, and make the ISO standards more user-friendly.

That being said, organizations will need to make adjustments to their QMS to meet the new requirements. The major impacts that organizations need to consider for ISO 9001:2015 certification include the following:

• Increased management responsibility
• Organizational identification of risks and opportunities
• Impacts of process implementation vs. guidance procedures
• Overhaul of internal audit requirements

Management Responsibility

The increase in management responsibility requires an organization’s objectives and targets to be:

• Business-driven
• More explicit in content
• Reviewed and monitored on a regular basis

Importantly, the QMS must be connected to the business strategy. This involves management taking ownership for the QMS and creating a vision and strategy for the organization, its employees, and customers to follow and interact with in a mutually beneficial manner. The idea is that this will foster a sustainable business plan.

Identification of Risks and Opportunities

The organization must identify and quantify the risks and opportunities presented by each new business endeavor or market driver they seek to enter. This will help management understand the full operational requirements and potential related consequences that must be addressed prior to moving the organization in a new direction.

The process of identifying risks and opportunities involves reviewing and evaluating employee skill sets, equipment capabilities, facility requirements, logistics requirements, environmental and safety risks, and others. In addition, quality control requirements must be reviewed in terms of possible training and equipment needs, and then verified as either adequate or in need of required changes prior to startup.

Process Implementation

The single largest change to the QMS is arguably the notion of written procedures guiding the organization vs. the use of a process approach to enhance the organization’s ability to exhibit systematic control over any/all changes to the products and/or services it provides. This change represents a shift in the approach regarding business operations.

Under a process approach, the management team must:

• Define inputs and outputs of each process
• Determine the correct performance indicator(s) to assure compliance and customer specifications have been met
• Assign appropriate responsibility for these steps

To comply with ISO 9001:2015, the organization must be able to stop a process and rectify the issues of concern prior to a nonconforming product and/or service being given to a customer. As such, employees are empowered to complete a root cause analysis and then notify management of possible change(s) required.

Internal Audit

Corresponding to the process orientation discussed above, the internal audit program will also need to be revamped to go from auditing a single clause to auditing an entire process. This may require additional auditor training for internal auditors, as well as an overall better understanding of the processes the organization follows in its daily business.

The following tips can all help modify the internal audit process to work under the ISO 9001:2015 standard.

1. Audit one complete process at a time. This will allow auditors to better assess the process itself, identify possible areas for review and improvement, and verify adequacy of current controls in place.
2. Develop flow charts that outline every step in the process(es) and the associated procedures, work instructions, and forms required to assure compliance of each identified step.
3. Look for areas throughout the audit where the product and/or service hand-off between departments and equipment cells may be unclear or confusing, leading to a potential nonconformance to the customer.

Big Steps toward Continuous Improvement

While any one of the changes discussed above would represent a significant improvement over the 2008 version of ISO 9001, taken together and implemented properly, the 2015 updates are set up to help organizations take large step towards continuous improvement.

Under ISO 9001:2015, day-to-day operations should:

• Be more functional and harmonious
• Allow for improvements in product and/or service hand-offs between departments
• Improve the consistency of delivering to the customer exactly what is requested
• Reward the organization with improvements to internal functions and lower costs over time

This is the first in a series of articles on food product recalls.

Salmonella outbreak in eggs; E. coli breakout in romaine lettuce. Both are getting a lot of attention right now. Unfortunately, no food company is immune to encountering situations like this that may lead to government warnings or a food product recall. Even plants with the best controls are at risk—human error, mechanical breakdowns, or sampling failures can happen at any time.

Growing Epidemic

The number and magnitude of product recalls has increased significantly in recent years. According to U.S. Department of Agriculture’s (USDA’s) Economic Research Service report entitled Trends in Food Recalls 2004-2013, the average number of food recalls between 2004 and 2008 was 304/year; the average number between 2009 and 2018 increased to 676/year.

Interestingly, the study does not cite riskier foods as the reason for this upward trend. Rather, the increase of food product recall events can be attributed to the following:

• An increasingly complex and global food supply chain system,
• Technology improvements in the detection of health risks, and
• Passage of two major food policy laws—FALCPA and FSMA—particularly related to the dramatic increase in undeclared allergen recalls.

Product Contamination Consequences

A full-scale recall involving food products can be detrimental to a food manufacturer or retailer. According to a survey conducted by the Grocery Manufacturers Association, 29% of companies that faced a recall within the prior five years estimated that the direct cost of the recall was between $10 million and $29 million—and that cost can be even greater when accounting for indirect costs.

There are three primary consequences of a major product contamination/recall event:

• Product recall expenses – product replacement costs, recall and redistribution expenses, product destruction costs, related crisis management consultation fees
• Business interruption – financial loss due to product unavailability, decontamination downtime, government action, brand damage, and loss of contracts
• Third-party liability – financial loss due to third-party property damage and bodily injury

In many cases, a recall event will result in decreased profits over the short run of 6-18 months. The long-term brand damage, however, can impact earnings over an even longer period. Given these trends and the potential associated impacts, every food business should be concerned with potential contamination risks.

Read:

• Part 2
• Part 3

The 20^th Annual Food Safety Summit, held earlier in May, proved once again to be an engaging and informative meeting for all in attendance. During the four-day event, food safety professionals participated in interactive sessions focused on food safety in the supply chain.

Throughout the Summit, Kestrel’s food safety experts observed several common themes and challenges that the food industry is facing — challenges that your business may be encountering today. Here are some of our key takeaways:

Key Topics

This year’s Summit featured a supply chain focus. Every company in the food supply chain—those that produce, handle, or distribute food-grade products/ingredients—has an obligation to its customers to provide safe and quality food. The Summit began with six certification programs related to food safety across the supply chain, including Preventive Control for Human Food, Foreign Supplier Verification, Professional Food Safety Auditor Training, Seafood HACCP, HACCP Training, and Certified in Comprehensive Food Safety.

The focus on food safety throughout the supply chain was discussed in lessons learned from recent food safety case studies, as well as in four afternoon workshops focused on departmental cooperation, traceability, effectively managing food safety, and global regulatory systems. In addition, the new Community Cafes provided attendees the chance to meet with the subject matter experts, including experts from Kestrel, for more in-depth conversations about topics focused on the supply chain.

Industry Trends

As is evident from the key topics covered at the Summit, the current trend is a much stronger focus on food safety within the global supply-chain, particularly coordination of various global regulatory systems, requirements, and agencies. Steve Mandernach of the Association of Food and Drug Officials (AFDO), Dr. Robert Tauxe of the Centers for Disease Control (CDC), Paul Keicker of the U.S. Department of Agriculture (USDA), and Stephen Ostroff from the U.S. Food and Drug Administration (FDA) participated in sessions on how the agencies are working together on food safety initiatives and the impact food safety. This included a conversation on whole genome sequencing of pathogens as a game changer for disease monitoring and response.

Ongoing Challenges

A major concern and ongoing challenge within the food safety community involves training regarding food safety responsibility across all levels of the organization. This not only includes large organizations with individuals who are directly responsible for food safety programs, but also smaller organizations where these responsibilities must be understood and followed by everyone at all levels. This corresponds to the need for food organizations to establish a much more comprehensive and effective food safety culture. The lack of this training, organizational responsibility, and overall culture is considered one of the primary causes of continued contaminated food outbreaks.

Regulatory Updates

In addition to the discussion regarding the use and benefits of whole genome sequencing, enforcement of FSMA and the expanding rules of Intentional Adulteration and Food Fraud Prevention Programs continue to be top areas of regulatory concern. Inspections and investigations related to supporting Food Safety Plans under these requirements will continue to expand, as evidence of compliance is required by all food organizations under the law.

Food Safety Summit 2019

Plans are already being made for the 21st Anniversary of the Food Safety Summit, which will again be held at the Donald E. Stephens Convention Center in Rosemont, IL from Monday, May 6 through Thursday, May 9, 2019. Mark your calendar. Kestrel looks forward to being an active sponsor and participant again next year.

Since it was launched in May 2000 following a number of major food safety scares, the Global Food Safety Initiative (GFSI) has aimed to “provide continuous improvement in food safety management systems to ensure confidence in the delivery of safe food to consumers worldwide.”

Recognized Schemes

GFSI is not a scheme in itself, nor does it carry out any accreditation or certification activities. Rather, a benchmarked scheme (e.g., BRC, SQF, IFS, FSSC 22000) is recognized by GFSI when it meets the minimum food safety requirements, as set out in the GFSI Guidance Document.

Strategy for Certification

Companies have the flexibility to choose which GFSI-recognized scheme they want to adopt, and can achieve certification through a successful third-party audit. Under GFSI’s concept of “once certified, accepted everywhere,” certification to any GFSI-recognized scheme is accepted by many international, national, and regional retailers and suppliers.

The following factors should be considered to ensure a successful GFSI strategy:

1. Adequate knowledge of the GFSI standards (e.g., BRC, SQF, IFS, FSSC 22000) and how they work within food manufacturing and packaging companies
2. Ability to use and implement document and records management and control
3. Training to implement the chosen standard and ongoing training in the standard/key program areas, including Hazard Analysis and Critical Control Points (HACCP) and internal audit
4. Meeting the building requirements of the GFSI standard
5. An integrated pest management system that meets the requirements of the standard
6. Dedicated role of a qualified plant sanitarian
7. A strategy that includes management commitment and allocation of budgets and resources
8. Proper management review meetings and records
9. Compliant food safety and security
10. A corrective and preventive action (CAPA) process that meets the requirements of the standard
11. Approved supplier programs
12. Control of non-conforming product through disposal
13. Change management and acceptance by the organization
14. Product specifications that meet the requirements of the standard
15. Sanitation and chemical control programs
16. Deviation and variance tracking, reporting, and response
17. Product and raw material storage
18. Food-level Good Manufacturing Practices (GMPs), operating prerequisites, and compliance
19. Calibration of measurement devices
20. Emergency response and contingency plans

The GFSI system provides a high degree of confidence that food safety management systems are adequately designed, implemented, and maintained. Certified companies tend to be more efficient and profitable and have more effective shared risk management tools for brand protection. Ultimately, certification results in improved consumer confidence, simpler buying, and safer food throughout the supply chain.

In recent years, companies have been generating vast and ever-increasing amounts of data associated with business operations. This trend has led to renewed interest in predictive analytics, a field which focuses on analyzing large data sets to identify patterns and predict outcomes to help guide decision-making. While many leading companies use predictive analytics to identify marketing and sales opportunities, similar data analysis strategies are less common in occupational and process safety. Although the use of predictive analytics is less common in the field of safety, the potential benefits of analyzing safety data are considerable.

Just as companies are currently using customer data to predict customer behavior, safety and incident data can be used to predict when and where incidents are likely to occur. Appropriate data analysis strategies can also identify the key factors that contribute to incident risk, thereby allowing companies to proactively address those factors to avoid future incidents.

Predictive Analytics: In Theory

Let’s take a step back and look at what predictive analytics is and what it does. Predictive analytics is a broad field encompassing aspects of various disciplines, including machine learning, artificial intelligence, statistics, and data mining. Predictive analytics uncovers patterns and trends in large data sets for the purpose of predicting outcomes before they occur. One branch of predictive analytics, classification algorithms, could be particularly beneficial to industry, especially when it comes to avoiding incidents.

Classification algorithms can be categorized as supervised machine learning. With supervised learning, the user has a set of data that includes predictive variable measurements that can be tied to known outcomes. The algorithms identify the relationships between various factors and those outcomes to create predictive rules (i.e., a model). Once created, the model can be given a dataset with predictive variable measurements and unknown outcomes, and will then predict the outcome based on the model rules.

Predictive Analytics: In Practice

Like many in the transportation industry, this railroad had experienced a number of derailments caused by broken rails. Broken rail derailments can have particularly severe consequences, since they typically occur on mainline tracks, at full speed, and with no warning of the impending broken rail. Kestrel was asked to create a predictive model of track-caused derailments on a mile-by-mile basis to identify areas of high broken rail risk so the railroad could target those areas for maintenance, increased inspections, and capital improvement projects.

Penalized Likelihood Logistic Regression

As described above, classification models learn predictive rules in an original data set that includes known outcomes, then apply the learned rules to a new data set to predict outcomes and probabilities. In this case study, Kestrel used a logistic regression modified by Firth’s penalized likelihood method to:

• Fit the model
• Identify eleven significant predictive variables (based largely on past incidents)
• Calculate broken rail probabilities for each mile of mainline track based on track characteristics

Final Model

The final model calculates a predicted probability of a broken rail occurring on each mile of track over a two-year period. The results suggest that the final model effectively predicted broken rail risk, with 33% of broken rails occurring on the riskiest 5% of track miles and 70% occurring in the riskiest 20%. Further, the model shows that the greatest risk reduction for the investment may be obtained by focusing on the 2.5% of track miles with the highest probability of a broken rail. This ability to predict where broken rails are likely to occur will allow the company to more effectively manage broken rail derailment risk through targeted track inspections, maintenance, and capital improvement programs.

Implications for Other Industries

The same general approach described in the above case study can also be applied to other industries—using KPIs to determine predictive variables and incidents as the outcome. The process is as follows:

• Measurements for defined variables would be taken regularly at each facility or unit. Precision increases as the measurements become more frequent and the observed area (facility/unit) becomes smaller.
• Once a sufficient number of measurements has been taken, they would then be combined with incident data to provide both the predictive variable measurements and the outcome data needed for training a model. This dataset would be fed into a logistic regression or other classification algorithms to create a model.
• Once the model has been created, it can be applied to new measurements to predict the probability of an incident occurring at that location during the applicable timeframe.

Once predicted incident probabilities have been found, management would be able to focus improvement resources on those locations that have the highest probabilities of experiencing an incident. The classification algorithms also identify which factors have predictive validity, so management will know how improving those factors will affect the predicted probability of incidents occurring. In other words, they will know which factors have the strongest relationship with incidents and can focus on improving those first.

Data-Driven Decisions

Industrial companies are generating and recording unprecedented amounts of data associated with operations. Those that strive to be best-in-class need to use that data intelligently to guide future business decision-making.

The versatility of predictive analytics, including the method described in this case study, can be applied to help companies analyze a wide variety of problems. In this way, companies can:

• Explore and investigate past performance
• Gain the insights needed to turn vast amounts of data into relevant and actionable information
• Create statistically valid models to facilitate data-driven decisions