Loss accounting is the process of capturing and categorizing production losses and their causes, and using this information to drive continuous improvement.
In the previous article, I defined lost production and introduced the choke model, which is an analysis to determine where production losses occur in the process and where to install short interval control (SIC). In this article, I explain the loss accounting process and how to use a loss accounting system to drive continuous improvement.
Essentially, loss accounting is the use of information on production losses to drive continuous improvement in overall equipment effectiveness (OEE), also known as asset utilization. Lost production is defined as saleable output that could have been produced but was not, for some reason. Obviously, this definition encompasses a vast array of possible causes. Rather than attempt to record and act on every possible event or cause, loss accounting is a process of quantifying and capturing lost production by assigning it to categories that can later be aggregated and effectively analyzed, so that problem-solving efforts can target the biggest and best opportunities.
Why is OEE so important in the process and energy industries? In large, capital-intense facilities, keeping assets performing at maximum capacity is the only way to make a return on investment in the long run. Once an asset is built, it rarely makes sense to run it at less than full rate, even when prices are low. This is due to the high capital cost and also fixed costs associated with operating them.
Since production loss accounting focusses solely on OEE, it must be emphasized that this is only one facet of manufacturing performance. Management is, of course, concerned about all aspects of performance, including health and safety, operating costs, sustainability, and regulatory compliance, which are all essential to long-term viability of the business and maintaining a license to operate. Nevertheless, OEE is a key performance indicator (KPI) that merits its own accounting system and continuous improvement initiatives.
The fire-fighting trap
To solve a problem, you need to be aware of it, you need to understand it, and you need to assign appropriate expertise and resources to solving it. This is especially challenging in large, complex operations, because of the large number and wide variety of problems. Some problems are infrequent but cause big losses. Others, are small but re-occur frequently.
Why do problems reoccur? Usually, for one of two reasons. Firstly, the organization is stuck in a "fire-fighting" mode, in which available resources are consumed by the immediate task of keeping the plant running, and time is not being dedicated to understanding root-causes of problems so that preventive measures are identified. For example, when the pump fails, the bearing is replaced and the pump put back in to service, but no-one investigates why it failed.
Secondly, many of the persistent problems are actually difficult to solve (otherwise, they would have been fixed long ago). Solving them will require something new, such as new information or outside expertise, or doing something different that hasn't been tried before. In many cases, it takes repeated iterations of trial-and-error to permanently solve problems. The problem-solving process—incident investigation, hypothesis forming, brainstorming solutions, testing/trialling, reviewing the results, and learning—may need to be repeated numerous times before a truly effective solution is found. Difficult problems take out-of-the-box thinking, time, and organizational tenacity to eliminate.
If everyone is focussed on today's problems, and then forgets about them tomorrow, the organisation will never break out of fire-fighting mode. The purpose of loss-accounting is to capture the critical information management needs to make the tough decisions that will make a difference and lead to change—such as assigning appropriate time and resources to eliminating the problems that cause the biggest losses, thus helping the organization to climb out of the fire-fighting trap.
Performance awareness
To get a sense of how well people understand current performance, try asking someone in your workplace what OEE is for the year to date, or for the current month, or this week. If you ask a production supervisor or plant manager, they should know. If they were in the daily production meeting, they should at least know what the previous day's production was compared to target and the causes of any losses. But are people aware of the cumulative impacts of different types of losses over time? Can you name the top 3 causes of lost production and quantify them (e.g. as a % of total losses)? If people aren't clear on the answers to these important questions, it's unlikely that the organization is aligned on production optimization priorities.
Without a robust loss accounting system place, and regular reviews of losses, it should be no surprise that we struggle to answer these questions. People may be completely unaware of the amount of opportunity that is lost. For example, small rate losses over the course of a year may be causing more losses than some of the biggest breakdowns that people can remember. It's difficult to know without a good accounting system. Worst of all, people may become accustomed to the status quo and believe that performance is "as good as it can be" or that losses are unavoidable, or to blame them on external factors beyond their control when in fact, most are preventable. Without readily-available facts, it's difficult to counter such beliefs.
Consensus
One of the goals and benefits of loss accounting is to build consensus across the organization. In the absence of facts, there may be disagreement on the causes of poor performance and thus on priorities and budgets. For example, there may be lack of alignment between the functional groups—operations, maintenance, and engineering. Management's perception may be divorced from the reality of day-to-day production problems. Process engineers may be working on projects to increase the capacity of certain equipment, when it's not the cause of losses. Perhaps the control engineer has come up with a great idea to reduce variability in part of the process, but if it is not contributing to losses at a choke point, what benefit will it have? Implementing loss accounting will dispel myths or false narratives, for example that most lost production is caused by operators, or by maintenance.
Building consensus and alignment is one of the biggest challenges in any organization. People have biases or 'pet projects' that they are invested in. At the same time, good projects may be ignored or fail to get buy-in from management because there aren't enough facts to prove the business case. Quantifying losses challenges paradigms about which problems are more important than others. Effective loss accounting leads to transparency and clarity on the causes of lost production, and thus focusses discussion and collective effort towards solving problems, not debating which problems to solve.
Of course, we would like to work on solving all the problems. But with limited time and resources, an organization cannot afford to work on every possible opportunity and tough decisions have to be made. That is why loss accounting is so important—the facts force people to collectively accept a common understanding of the status quo, the gap compared to achievable potential, and the knowledge of where to focus efforts to deliver real, measurable results. Increasing OEE is a goal shared by all departments—operations, maintenance, engineering, and projects. In that sense, it can be a great way to unite the organization in its common pursuit.
Recording and categorising losses
Like OEE, production losses are divided into three main types, downtime losses, which are periods when the production rate was zero because equipment was not available to operate, rate losses, which are the difference between actual production rate and maximum achievable rate when the plant is running, and quality losses, which are any produced output that did not meet the required specification.
Each of these categories may be further broken down into sub-categories. For example, downtime losses may be grouped into planned maintenance, breakdowns, emergency shutdowns, power outages, ... etc. In fact, a rigorous loss accounting system may consist of a hierarchy of categories and customized codes to capture all manner of useful information, such as the equipment item in the asset hierarchy, the type of failure (e.g. flange leak, seal failure, bearing failure, ...etc.) and the causes, once they have been established.
Recording losses is time-consuming but the idea is to record the essential information to analyze losses later and identify the best opportunities to improve. For example, if you think that impeller failures might be a common problem, you should be able to go into the loss accounting database and search for "impeller failure" and get a list of loss events, and the total production losses associated with them in a given time period. You should be able to pull up data on the most common root causes of these failures and from that identify potential preventive measures.
Most of the useful information needed to understand losses cannot be deduced at a later date by analyzing production data
Despite what vendors of digital technology may claim, most of the useful information needed to understand losses cannot be deduced at a later date, simply by analyzing production data or maintenance records. Plant information systems (i.e. process historians) record a lot of process measurements and equipment conditions but they do not capture important details and contextual information about the operation. Maintenance records tend to focus on what work was carried out, not why the equipment failed.
Control room operators, supported by a team of field operators, have a deep situational awareness based on information they acquire from a variety of diverse sources, including senses (visual, aural, smell, touch), ongoing communications with maintenance, and their awareness of the state of the process including any abnormal equipment conditions or ongoing maintenance activities. Based on this awareness and a rational mental model of the dynamics and causal behaviour of the process, they are in a good position to record relevant details, including their interpretation of the cause of the loss. Capturing this, along with input from supervisors, technicians, and engineers at the time of an incident, is essential to ensure the loss accounting system contains enough meaningful details to analyze and act on losses at a later date.
Recording losses is, of course, just the first step. Unfortunately, some organizations get this far and stop. When people are expected to spend time entering information into a computer system but nothing ever comes of it, they soon become disillusioned with the process and the quality of the information deteriorates to the point of being useless.
Loss reporting
To ensure the quality of loss accounting data and to ensure it ultimately leads to actions, losses should be reviewed daily, weekly, monthly, and annually by operations management. Consider the daily plan-control-report cycle in Figure 1 below. I highlighted the elements that are touched by loss accounting. As explained in a previous article, losses are initially captured by the central control room operator during the hourly short interval control (SIC). At or near the end of the shift, the shift supervisor reviews the SIC with the operator and enters the losses into the system (or approves them if the operator has already done so). They are then summarized on the shift report, and should be part of the handover to the next shift. A summary of the losses over the previous 24-hours are added to the daily-weekly operating report (DWOR) and reviewed in the daily production meeting.
The daily production meeting is the first time when all the relevant discipline leads (operations, maintenance, engineering, health, safety, environment, etc.) have a chance to review the previous day's production losses as a team. The meeting should have a clear agenda and one of the first items, after reviewing any safety and environmental incidents, should be to review production compared to target and the daily losses. This is an opportunity for people to question the facts and figures, ask for and share additional information, and establish a consensus on what actually happened, or at least develop some hypotheses, and determine what immediate action is needed (if any).
The key principles at work here, are that the aggregated facts and figures from the previous 24-hours are reported to the next level of management by the responsible person who has the full picture of the events that occurred—i.e. the production supervisor—and the facts are available when the key decisions are made. The bare facts on actual production vs. target, and the total losses in terms of availability, rate, and quality, have limited meaning on their own and are insufficient to understand what happened (this explains why posting KPIs or 'dashboards' on a wall in a corridor has very little impact).
Effective decisions are made in regular formal meetings with the right people, at the right place, at the right time, with the right information. With the facts on the table, the responsible managers can efficiently dig deeper to analyze what happened, by questioning, challenging, and discussing the broader context and details of events, and thus agree on an appropriate action (the meeting has an action log). For example, other than corrective actions to deal with the immediate consequences of a production loss, a decision can be made at the daily meeting on whether further investigation or fact-finding is necessary to complete the loss report, and whether a root cause analysis (RCA) should be initiated (a similar process exists for safety incidents).
Recall the concept of mental models, which I touched on previously. What's happening in the daily meeting, is that two parallel information systems are being updated—one is the loss accounting system, which is a computer database containing factual information on all the losses, the other is the mental models of the management team, which are much richer and more sophisticated representations of reality based on intelligence and experience. Thus, just like the hourly SIC by the control room operator, the discipline of daily review of OEE and production losses ensures that the mental models of the management team are kept aligned and up to date with reality.
The focus of the daily meeting is too myopic to solve problems—it is to review the previous 24 hours and decide what immediate action is needed to remain on track to achieve the weekly production targets. Therefore, a similar process of aggregating the information is necessary at the weekly, monthly, and annual review frequencies. At each level, the aggregated cumulative facts are reviewed for the current period so that appropriate action can be taken by the right people. At the weekly operations team meeting, a summary report of the losses over the previous 7 days is reviewed to identify significant issues that need addressing immediately. Then, usually at a bi-weekly or monthly frequency, a dedicated team of specialists reviews a summary of the latest losses and decides whether current improvement initiatives are delivering as expected. The management team, may direct this team to focus on certain losses and come up with new improvement ideas and initiatives. Finally, year-end is a good time to do a detailed analysis and summary of all losses over the previous 12 months. Figure 3 below is an illustration of a waterfall plot, which is commonly used to summarize annual total losses.
The purpose of this article, and that of loss accounting, is not to explain how to solve problems—that is another topic entirely. Rather, it is to help organizations determine which problems they must solve to increase production and OEE, and to ensure they make the tough decisions needed to do so. If all your production losses are due to one known cause, loss accounting is not going to have much value—it will only tell you what you already know. However, large organizations responsible for complex production processes can easily become overwhelmed by the number and diversity of problems. One of the outcomes of loss accounting may be that you simply stop working on some issues in order to focus on a smaller set of problems that will yield greater results.
Once you do start working on solving problems, loss accounting gives you automatic and ongoing feedback on the results. This is important because most production issues are inherently difficult to fix—it may take weeks, months, or even years of problem-solving efforts before they are eventually eliminated. Like SIC, loss accounting and loss reporting is the foundation on which to build a learning organization.
In summary
What are the benefits of loss accounting?
Awareness and consensus on current performance and the causes of lost production
The facts needed to allocation time and resources to solving problems
Organizational learning and tenacity to solve difficult problems
Systematically eliminate losses and escape the fire-fighting trap.