Going rogue has been saved
Deviation from standard work processes is common. But rather than view process deviation as the enemy, companies can use it to drive positive outcomes by applying behavioral principles in the workplace.
It is Sunday afternoon, and you are enjoying watching your favorite quarterback at his finest. Your hometown football team is down 34–28, with the ball on the opponent’s 36-yard line. The clock is winding down, and it is 3rd and 2. With the game on the line, the quarterback calls “Omaha, Omaha,” and the center snaps the ball. The offensive coordinator had called a jet sweep to the star wide-out. The quarterback, however, well known for shrugging off the play call, drops back, waves off the wide-out, and instead throws a Hail Mary pass into the end zone. This time, the quarterback’s decision to deviate, to ignore the offensive coordinator’s instruction, is heralded as a brilliant game-winning decision.
You are thrilled at the outcome, but you have to wonder, why didn’t the quarterback run the play for an easy pickup of two yards as he had been directed to do? Did he see something in the defensive line that made him doubt the play would work? Did he see an open route that prompted him to throw the ball, or did he simply decide to deviate from the instructed play for some other reason, or for no reason at all? You’re not the only one wondering, either. Why the quarterback went rogue will be asked and answered at the post-game press conference and shown over and over on every news and sports telecast for days to come.
Lest you think professional football players are the only employees to occasionally throw out the playbook, the evidence suggests otherwise.
Lest you think professional football players are the only employees to occasionally throw out the playbook, the evidence suggests otherwise. The research we present in this article suggests that workers will quite readily disregard standard business processes when the incentives or contexts that they face encourage them to do so. These studies argue that the urge to deviate persists over time and across business and other operational settings. Many times, process deviations are negatively perceived, but perhaps just as frequently, they are viewed as essential drivers of superior performance.
The cumulative effect of individual decisions to adhere or deviate from standard process is hard to quantify. Given that employees’ actions are a response to their environment, maybe it is most important to understand why they do what they do. Imagine the power of having the kind of hindsight that football coaches enjoy on Monday morning. What if businesses better understood why employees deviate from approved processes, even when their actions may have serious consequences? Or when employees are forced to deviate due to a system failure, what can be learned from their response? Can businesses better manage the occurrence of rogue behaviors and use these behaviors as critical input to a process of continuous improvement?
In order to inform this discussion, we offer the perspectives gleaned from three related studies of work process deviation. The research that we describe takes place in a variety of contexts, including enterprise resource planning (ERP) implementation, travel and hospitality management, and health care. This breadth of context illustrates how pervasive issues of process deviation can be and offers important insights for managers in pursuit of improved operational performance. In support of that end, we formulate and offer research-based recommendations for managing individuals and their actions within a complex and multifaceted operating environment.
Behavioral economics is the examination of how psychological, social, and emotional factors often conflict with and override economic incentives when individuals or groups make decisions. The field has its roots in the work of Nobel Prize winner Herbert Simon, who as far back as 1959 questioned the classical economic theory that individuals rationally maximize the outcomes they seek (that is, their “utility”) when making choices.1 Ever since, scholars have argued, and demonstrated, that in the face of uncertainty, humans employ all manner of simple, easy-to-use, but often inaccurate cognitive shortcuts. People are not maximizers but rather muddlers, struggling to cope with a reality that is much less certain, more complex, and more variable than they admit. Worse, as has been shown repeatedly by another Nobel Prize–winning behavioral economist, Daniel Kahneman, “We can be blind to the obvious, and we are also blind to our blindness.”2
The implications of cognitive limitation and bias in decision making are as varied as they are vast. Consequences stretch across industries and applications. Managers therefore need to understand something about those aspects of human perception and cognitive processes that can lead us astray. In short, they need to understand something about the psychology of choice.
This paper is one in a series offered by Deloitte that intends to address this need of managers attempting to improve the performance, growth, and innovative capabilities of their organizations.3 Each paper in this series examines the influence and consequences of behavioral principles on the choices people make related to their work. The reader can expect to see some common themes emerge that are related to the framing of choice, the uncertainty of outcomes, the influence of time (and timing), and a general inability to effectively process information. Each paper will provide a set of practical guidelines for recognizing, managing, and/or mitigating the effects of these very natural human limitations and biases in an applied setting.
Collectively, these papers aim to illustrate how an understanding of biases and cognitive limitations is the first step in developing countermeasures that limit their impact on the organization. Alternatively, they may help organizations and individuals identify strategies that ethically exploit these inescapable elements of our humanity to the benefit of a company’s performance, growth, and innovation.
Standardization of business process has long been viewed as a benefit to organizations through its ability to promote the deployment of process knowledge within and across operations, as well as to enhance the ability to measure and improve operations.3 Implicit in this perspective is the understanding that the information technology and protocols that support a business process embed a view of how the process should best function, along with a set of assumptions about the context within which the process will unfold. These systems often cross the boundaries of subunits within the business (as well as boundaries between businesses). That means that workers must make choices about whether the prescribed manner of completing work is actually the best way to do things, or whether they should deviate from the prescribed approach and act based on their own assessment of the situation.
Research suggests that not only is process deviation embraced by many, but also the intention to deviate, even if thwarted, will persist over long periods of time.
Our first study looks at how persistent employees can be while pursuing the opportunity to deviate from the prescribed process when they believe it does not properly support actual business need. It does so in the context of the kind of ERP systems implementation that companies are known to spend tens or hundreds of millions of dollars deploying. The research suggests that not only is process deviation embraced by many, but also the intention to deviate, even if thwarted, will persist over long periods of time.
Where the first study examines the intrinsic motivation of workers to deviate based on a perception of “mismatch,” the second study evaluates the impact of extrinsic motivation in the form of measurement and incentives. Here participants in the hospitality industry are presented with a revenue management (RM) system that uses sophisticated algorithms to balance supply, demand, and time to prescribe pricing for hotel rooms. Users have the ability to deviate from these recommendations based on their own judgment and sense of conditions that might not be fully accounted for by the established technology and process. The study addresses the question of how managers can influence worker behavior related to process deviation through the use of measurement and incentives. The evidence suggests that the choices managers make in this regard can matter a great deal.
Beyond establishing that the motivation to deviate from standard process, particularly as instantiated in IT, can be both internal and external, with predictable results, findings from the first two studies also suggest that such deviations may not always be a bad thing. In fact, they might represent actual opportunities to “do better.” In particular, the RM case explicitly recognizes that workers may have information external to the “algorithm” that can inform better decision making.4 But if this is so, what is the consequence to ongoing learning? If workers routinely deviate (for better and worse) to adjust to the reality of a particular situation, does this lead to overall organizational learning and improvement?
Our third study suggests that the hurdles here may be substantial. This study’s context is health care: The researchers followed nursing professionals to observe points at which they were required (in their opinion) to deviate from standard process in order to address an emergent situation. In this context, the drive to deviate from established process is framed as “operational failure” along with the need to address a situation for the benefit of the patient (as opposed to the transaction). Regardless, as before, most events were instigated by a real or perceived failure of information flow (and in some cases material flow) between organizational units (that is, the process did not embed the appropriate information). Also, as before, we see workers (here nurses) handle these situations in due course.
The unique focus of the third study was whether and how information about the process failure was communicated to the rest of the organization and whether it was incorporated as a learning opportunity into the general process. It seems that, in most cases, such was not the case. In particular, the research found that:
“Highly interdependent front-line workers do not control organizational processes responsible for the majority of failures they encounter and have a difficult task convincing powerful associates that these obstacles warrant solution efforts, making it likely operational failures will persist.”5
The implication of this finding and the findings of the other research are substantial and will be discussed following a deeper review of each study.
“I know better!” In its simplest terms, the first study helps to explain why employees choose to work around a newly installed ERP system, deviating from carefully designed and standardized protocols and processes. Employees are motivated to behave this way because of their perception that the system will not work for them or let them get work done their way. At its core, their motivation is akin to the hypothesis that our quarterback deviated from the play called because he believed it wouldn’t work. In short, he thought he knew better.
The first study examines intended business process deviation following ERP deployment. An organization typically implements ERP to standardize work processes and promote greater consistency.6 However, achieving consistency of outcomes is predicated on workers choosing to adhere to the system-embedded business process that is defined for them, including following prescribed steps for completing work. It is possible, however, that system designers’ perspectives vary by the individual on the proper balance between processing “needs” and “prescribed” procedures.7 Furthermore, other research recognizes that some workers are more naturally willing to deviate from prescribed approaches than others.8 This is a critical issue, particularly when a person is faced with a variety of processing options.9
In the study, researchers used a controlled experiment to influence worker perceptions of how an ERP system deployment “fits” with the needs of the business over the short and long term (see the “Study design” sidebar for a more detailed description of the research approach).10 It tested whether employees’ intentions to deviate from standard processes following a large-scale systems implementation are contingent on perceptions of fit and the ease with which one can deviate from the process. In doing so, it also examined the durability of those intentions over time.
Researchers used controlled manipulation of context using four alternative scenarios outlined below. The scenarios varied the extent to which the standard business process represented a good “fit” for the business need (vertical axis) and how easy it was to deviate from (the authors use the term “circumvent”) the process in order to react.
Four case descriptions were developed to see if the reactions predicted by the research framework would be reflected in behavior. One of the four cases was presented to each of the 335 managers enrolled in an evening MBA program.
After reading their specific condition’s case study (for example, low misfit and high ease of circumvention) respondents answered questions focusing on their short- and long-term intentions to deviate from the prescribed business process.
Average intention to deviate was compared across each of the cases and subjected to rigorous statistical testing and identified significant differences between each of the test groups. In short, the intention to deviate from prescribed business processes was significantly influenced by the level of perceived fit, as well as by the ease with which such deviations could be accomplished.
Employees’ intentions to deviate from standard process depend on how strongly they perceive there to be a technology misfit and the apparent availability of ways for them to circumvent the imposed system protocols. Importantly, workers appear to be resilient in their intentions when ways to work around the system are not readily available. Workers demonstrate patience as they search for circumvention pathways, initially complying with system protocols. In the frequent cases where the organization winds down its employee adoption efforts, the tenacious intention of such employees eventually manifests itself in process-deviation behavior.
This first study further suggests that there are some types of employees more likely to accept changes imposed by enterprise system implementations (that is, to resist the urge to deviate). The findings show that employees with supervisory experience intended lower levels of long-term process deviation. Related research argues this is because supervisors often demonstrate a greater understanding of the interdependencies that work settings require and are therefore more willing to experiment with technologies that may promote greater collaboration.11
Based on these findings, organizational success with systems implementation seems to depend on long-term employee buy-in and adoption throughout the implementation life cycle, with an emphasis on late-stage monitoring. Organizations are wise not to be fooled by the emergence of a seemingly new normal, remembering the adage “old habits die hard.” Study participants demonstrated a willingness to deviate from prescribed business process, over the long term, when they believed it was necessary in order to meet a better-understood business need.
Communication of interdependencies, within and across business units, represents potentially important strategies for minimizing these deviations. However, for those employees more resistant to change, the next study illustrates the power of incentives to motivate behavior change.
“It’s in my hands!” What if instead of punishing employees for acting independently, organizations allowed their employees to exercise discretion and even rewarded them for doing so in a competent manner? Back to our football metaphor: What if the incentive system the quarterback was motivated by influenced his willingness to change the play called, as well as his subsequent performance? In this second study, researchers show how providing employees with the autonomy to throw out the playbook can result in improved performance, as long as certain organizational conditions are in play.
In the hospitality industry, hotels often rely on RM systems to provide booking agents with room pricing recommendations. These systems offer recommendations that are statistically likely to maximize revenue given the number of rooms available throughout the booking period.12 Hotel chains vary in the level of adherence to these recommendations that they expect from workers.13
Many large hotels empower agents to exercise judgment when deciding whether to reject bids for available rooms—in other words, to deviate from RM-suggested prices. This discretion allows agents to incorporate information that may not be accounted for in the RM algorithm (for example, another conference in town, unique weather conditions, local construction).
The authors of this second study endeavored to identify those factors that influence agents’ willingness to “throw out the playbook,” deviating from RM recommendations. Specifically, how do performance measurement and employee feedback impact their booking decision making?
In this study, key performance metrics (KPIs) took two different forms. For the first, “accept errors” were tracked, wherein an agent accepts a bid below system guidelines, often due to diminishing time in the booking period. A second KPI was measured as the difference between revenue generated and maximum possible revenue (based on suggested pricing). The KPIs were designed to return basically the same information, but they were intentionally framed differently (focus on benefit versus penalty/cost) to isolate any potential stress-induced psychological responses (see the “Study design” sidebar for a more complete description of the research).
In this study, 120 experienced hotel sales agents were exposed to a RM system that presented them with pricing recommendations under varying conditions.
Each agent’s task was to allocate a fixed number of rooms to prospective customers over a fixed period of time. Rooms that were not booked by the end of the booking period had no monetary value. The agents received incoming bids and decided to allocate a room and generate revenue, or reject the bid in the hope of selling to a higher-paying customer before the time expired (at which time unsold rooms would have zero value).
Agents were divided into three treatment groups. Each group saw the same pattern of bid prices and arrival times, but they differed in the kind of feedback they received on their performance.
A control group received no performance feedback.
A penalty-framed group was offered continuous performance feedback that emphasized missed opportunities caused by deviations from the RM recommendation (especially bids accepted that were below the RM recommendation).
A benefit-framed group was offered continuous performance feedback that emphasized situations in which the organization benefited from an RM deviation (especially when deviation resulted in a subsequent higher-value sale).
By construction, the feedback measures were highly correlated, meaning that differences in behavior were likely related to the perception of the performance measure and not the information conveyed by them. In addition to overall performance, study participants were also monitored for the stress levels they experienced. This information was also correlated with performance by the researchers.
The study demonstrated that reservation booking decisions were heavily influenced by the feedback conveyed. Even with the shared objective of revenue maximization, feedback form had an observable impact. Specific findings suggest that:
Based on this second study, one could argue that the motivation to deviate from standardized business processes and the systems that support them does not spring simply from the perception that the process is not sufficient for the task at hand (as in the first study) but can also be enhanced by the external metrics that are applied to workers. In this case, the framing of deviations as positive or negative increased or decreased the rate of deviation exhibited by workers. Furthermore, the study illustrates the impact that negative framing has on stress levels of workers—a factor that managers might rightly consider.
Coincidentally, the study also demonstrated an instance in which higher levels of deviation are actually associated with higher levels of performance, suggesting that such deviations are not always a bad thing for the business. One trick, it turns out, comes in understanding how difficult it can be for organizations to capture learning from process deviation in order to standardize and improve overall performance. Such is the topic for the third study we examined.
“I have no choice!” Consider, for a moment, that instead of choosing to deviate, an employee feels compelled to do so because of an operational breakdown. In this third study, researchers examine how health care workers respond when faced with operational failures, improvising to protect patient care—most akin to the possibility that the quarterback saw a breakdown in the play called. With no other choice, he had to throw out the playbook, react to the offensive breakdown, and improvise in pursuit of a better outcome.
Operational failures regularly occur in complex service organizations, including health care. In addition to wasting valuable resources and subjecting patients to unnecessary procedures, failures often decrease employee morale and increase thoughts of quitting.18 More concerning, even seemingly small failures may result in a threat to the patient’s well-being.19 This tends to happen when problems delay patient care, compromising patients’ health and creating incremental work for nurses. These problems also can make it more difficult for nurses to remember and manage all the necessary steps of patient care and thereby increase the potential for errors.20 This represents a critical issue given estimates that medical accidents kill more people each year than do motor vehicle accidents, breast cancer, or AIDS.21
A group of 26 nurses at nine hospitals were shadowed by researchers for nearly 240 hours. During this time, detailed work activity information was recorded.
Following observational data collection, researchers conducted interviews with half of the nurses. The semistructured interview protocol was developed to understand how operational failures impacted nurses’ productivity and their sentiment about their work-related roles and caring for patients.
Operational failures were classified as either problems or system errors. Problems included events such as something needed by the worker being unavailable or even present in a way that interfered with task completion. System errors included situations where task completion was deemed unnecessary or incorrect. System errors were viewed as errors in the flow of work that may or may not have resulted in a medical error (for example, an unnecessary step may not cause harm to a patient, but it still wastes hospital resources).
The impact of each problem on the overall system was quantified based on multiple dimensions of performance, including number of extra tasks required, direct and indirect time spent on the problem, number of interruptions, direct and indirect delay of task completion, patient risk, number of people involved in problem resolution, and waste. The approximate cost of each problem was estimated using these metrics along with estimates of nursing wages, hospital costs, and costs of medical errors.
Often, operational failures stem from failures in the supply of materials or information across organizational boundaries (different departments), with employees quickly compensating for their occurrence. As such, the total costs incurred for these issues often go unseen, and the issues therefore are challenging for organizations to learn from and improve upon. Learning from these required process deviations is thwarted because workers (nurses in this case) quickly execute work-arounds and then move along to the next task, with no organizational effort aimed at capturing and incorporating lessons learned.22
Learning from these required process deviations is thwarted because workers (nurses in this case) quickly execute work-arounds and then move along to the next task, with no organizational effort aimed at capturing and incorporating lessons learned.
In our third study, researchers seek to offer prescriptions for reducing operational failures by first understanding their fundamental nature. Operational failures are viewed as common in hospital nursing units because interdependency and uncertainty are characteristic to these work environments.23
Task interdependency can compel nurses to depend on others, such as doctors, pharmacists, and janitorial staff, to work effectively. Furthermore, providing care to patients is ripe with uncertainty, with each patient having a unique condition, medical history, and response to treatment. Because of these two conditions, on occasion, nurses find themselves without the materials, information, or services to treat patients.
The study observed two types of operational failure: problems and system errors (see sidebar “The high impact of failure” for a more detailed description of the research). The distinction between these types of failures is meaningful to the potential for organizational learning and problem solving. Problems are defined as those failures that workers are aware of as they happen—these prevent nurses from completing a current task. In contrast, system errors are identified after the fact, as nurses realize their work was unnecessary or unproductive. These two types of failure differ in whether problems can be corrected or recorded as they occur.
A failure’s source was also considered by the researchers as another factor relevant to learning and problem solving. The study classified the origin of each problem as either internal or external to the nursing unit. This determination was deemed relevant to understanding the likelihood that the nursing unit could affect a solution to a given problem by itself.
Finally, the response of nurses and hospital managers to failures was recorded to determine if the failure was perceived as routine (a recurring or normal part of nursing) or warranting special concern. The distinction is important because routine failures represented deviations from standard process that were typically unanalyzed and therefore not subject to systemic learning or correction (even if they solved the problem at hand).
Failures were ranked based on three dimensions, including impact on patient care, complicated resolution, and interaction with the hospital beyond the nursing department.
Eleven percent of failures had a significant impact on patients, nurses, and hospitals. These failures required an average of 5.3 additional tasks, 11.6 minutes of nurses’ time, and assistance from two additional people. In comparison, low-impact problems required approximately two minutes to resolve, and no extra tasks or people.
High-impact failures negatively impacted patient care efficiency, with an average delay of 34 minutes. Delayed tasks tended to include time-sensitive issues such as lab tests, medication, and meals. The efficiency of patient care was not typically affected by low-impact problems.
High-impact failures were also associated with hospital waste within and beyond the specific nursing unit. For these failures, the average task was not completed until 42 minutes after the failure was fixed and nurses were disrupted 1.7 times.
The most serious consequence of high-impact failures was risk to patient safety, with a mean risk of 1.8/3.0 (0 = no risk, 1 = patient discomfort, 2 = potential for risk if other factors present, 3 = failure itself could cause harm).
Overall, individual failures costs were estimated to average $414, and ranged from $16 to $1,708.
Low expectations and lack of control contributed to the acceptance of failures.
Despite the difficulty of revealing the occurrence and cost of organizational failures, they are worthy of the effort to understand and adapt standard processes to eliminate them. Over time, the costs of these failures can be enormous.26 From this study, we conclude that workers often react to operating failures in a competent manner, deviating from standard process in order to accommodate circumstances and information that have not been anticipated by process designers. The research establishes that although these impromptu responses often accommodate the situation or problem at hand, they fall short of realizing the opportunity to trigger learning, adaptation, and better long-run performance. In this case, process deviation seems to be a positive, if unrecognized, aspect of the overall system.
Taken together, these studies have shown that deviations from standard business process are common, are driven by a complex set of employee motivations, and result in a continuum of outcomes. Some employees will exercise their discretion and deviate from established guidelines in pursuit of performance optimization, as we saw with the hotel booking agents. Others will deviate when suboptimal working conditions prevent them from getting their work done, taking the initiative to find a suitable work-around that enables them to complete the task at hand. Unfortunately, deviant behavior can also be the result of more self-serving ambitions, as we saw with the employees who bided their time and actively worked to circumvent technology implementations they perceived as too rigid, inappropriate, or inefficient.
If one accepts that process deviation in all its forms is likely, how can managers increase the propensity for positive outcomes? Below we consider a three-step strategy—reduce, embrace, and leverage—that managers can employ to do just that.
During football season, we have the opportunity to see how organizations can translate well-managed rogue behavior into tangible success through these three steps discussed above. First, our favorite quarterback reduces ambiguity through his legendary understanding of the finer points of offensive and defensive systems and how they work together. Second, his greatest immediate incentive for positive deviation is the satisfaction of victory (and long-term financial compensation). And finally, 24 hours later the coaching staff is able to leverage the learnings of his rogue behavior in the film room to ensure continued success next week.