Police Performance Management in Practice:
Taking COMPSTAT to the Next Level

by William J. Bratton and Sean W. Malinowski

EDITOR'S NOTE: This article written by Lieutenant Malinowski and the Chief of Police, "Police Performance Management in Practice: Taking COMPSTAT to the Next Level" has been published as an OPINION piece by the Oxford University Press, "Policing: A Journal of Policy and Practice" in the United Kingdom. By measuring the performance of police managers whilst holding them to account for crimes, they explain the role COMPSTAT played in fighting crime in these areas and look forward to see how police can continue to innovate and expand upon existing police performance measures.

Police Performance Management in Practice:
Taking COMPSTAT to the Next Level

by William J. Bratton and Sean W. Malinowski

September 9, 2008

Introduction

Much has been written about police performance management in the last decade. That, in and of itself should be considered a victory for a profession that, before the dramatic crime reductions of the 1990s, was almost universally discounted by the academic community about its ability to reduce crime and control behavior. If that were the only victory, we would all agree that it would be a hollow victory at best. So, what do we have to celebrate? What do we have to scrutinize even further and what can be done to capitalize on the tipping point effect on crime reduction that we have driven beginning in the early 1990s?

We can all celebrate the fact that due to our collective efforts, crime has been significantly reduced on both sides of the Atlantic (Nicholas et al., 2006/2007). Our insistence that ‘Cops Count’ is supported by the fact that when properly led, well trained and sufficiently equipped, the police can modify the behavior of the criminal element in our communities. Beyond the saving of lives and the reduction of crime and its victims, we can also appropriately take credit for helping to inject new quality-of-life initiatives into our inner cities, for spurring economic development and for returning large tracts of the urban landscape to the lawabiding.

By way of example, in Los Angeles, the once majestic MacArthur Park in the Rampart area had been overrun by thugs and dope dealers since the late 1970s. Gangmembers dominated the landscape and the local community suffered not only the loss of their park, but the regression in quality of life that comes from local businesses and jobs fleeing for safer environments. By intensely focusing our limited police resources (in 2007, the LAPD policed the city with a ratio of about one officer for every 445 residents, less than half the rate of New York City. With current staffing levels at 9,572, the bulk of the praise for the success of the crime declines must go to the men and women of the LAPD), holding police commanders accountable for crime reduction and by fostering creative solutions through the use of technology and the development of strong community policing partnerships, this park has now been returned to the residents. Parents take their children to play and pensioners meet and greet each other and play dominoes,where the drug dealers and gangs used to ply their trade. Legitimate businesses have returned and have spread quickly north down the corridor from Rampart toward Hollywood and south down Wilshire Boulevard toward the revitalized downtown.

Crime reduction and performance management

Our crime reduction success in Los Angeles (Part I Crime is down 35.5% since 2002) as well as the continued success in New York City is due, in significant part, to Performance Management. Performance Management has been described as

The creation and implementation ofCOMPSTAT, with its emphasis on measuring the performance of police managerswhile holding them accountable for the crime that occurs in their districts, was seen as revolutionary in 1994 when I was appointed New York City police commissioner. Now it is nearly universally viewed simply as the way we do business. What I will discuss here, in more detail, is: how performance-based management contributed to our success in fighting crime in New York and now in Los Angeles, how we can continually innovate and creatively expand police performance management through the development of new and improved computer assisted metrics, and how we can use predictive methods to create even more timely and successful intervention and crime reduction initiatives.

New York

When I became police commissioner of the NYPD in 1994, the city government was sluggish and the incidence of crime had actually begun to slow, due to the introduction of a somewhat watered down version of community policing. However, the city was still paralyzed by fear and the police department was still viewed as ineffective in dealing with crime and disorder. The leadership of the department was focused almost exclusively on corruption control. The prevailing notion was that no one ever lost their job over crime going up, but plenty had seen promising careers cut short over corruption issues among their subordinates. The conventional wisdom at the time held that the police were there to respond to crime as quickly as possible when summoned via the 911 system---the performance measure of the day being ‘response time’. In the case of major crimes, a very detailed and rigorous post-incident criminal investigation was conducted to identify the perpetrator in hopes of arresting same, and by doing so reducing that offender’s ability to re-offend. A fairly simple approach---random patrol, rapid response and reactive investigation---all while keeping an eye out for indicators of corruption, was executed, though the focuswas misplaced on responding to crime, not preventing it.

So, what changed?While there are any number of explanations that have been offered and staunchly defended by academics as to what really drove crime down in New York City and initiated a domino effect across the country, I was there and here is what I contend happened. For most of the period of the 1960s to the 1990s, many of the most influential politicians, researchers, the media and even some well-intentioned police leaders sought to limit the role of the police to first responders rather than that of first preventers.We were told that the causes of crime were economic, social, demographic and ethnographic and that we could have no impact on these so-called causes. Rather, we were encouraged to focus on response to crime and to measure our success by arrest numbers, clearance rates and response time (Bratton, 2006). In effect, we were not held accountable for doing anything to prevent crime. I, along with a number of police leaders at the time, did not accept this notion that the police could not modify behavior and control crime. Our introduction of a new system of management now known by the acronym for Computerized Statistics (COMPSTAT) was remarkable at the time, for its scope, speed of implementation, and its impact on performance.

COMPSTAT and NYPD

The development of the COMPSTAT system of police management involved not only a focus on measuring outcomes but also on managing for improved outcomes. Shifting from a passive posture to a focus on outcomes and on controlling behavior rather than just measuring our response, showed definitively, that after 40 years of uncontrolled crime increases, fear and disorder, in the 1990s we finally got it right. We, the police, had helped to create a huge and positive impact. We did it by focusing on and prioritizing the desired outcomes---less crime and more safety---and we began to achieve historic crime reduction and improved quality of life. Our new focus remains primarily on measures of effectiveness, not just activity and response. Crime may go up or down to some degree when influenced by many of the old so-called causes (that I prefer to describe as influences), but the quickest way to impact crime is with a well-led, managed and appropriately resourced police force that embraces risk taking and not risk adversity and a policing structure that includes accountability-focused COMPSTAT management principles, broken windows quality-of- life initiatives and problem-oriented community policing. In sum, you can expect that which you inspect. In the 1970s and 1980s we were inspecting the wrong things with the resultant failed outcomes.

Through a process of reengineering based on the continuous improvement of performance-based benchmarking and the adoption and institutionalization of best practices, the men and women of the NYPD drove some of the most dramatic crime reductions observed anywhere in history. The results of focusing on accurate timely intelligence, rapid deployment, effective tactics and relentless followup were indeed, dramatic. From 1993 to 1998, New York saw a 53% drop in the burglary rate, a 54% drop in reported robberies and a 67% drop in homicide (O’Connell and Straub, 2007).

Los Angeles

It is important to note, however, that one size does not fit all. The process as it played out in New York City was very different from the way it now works in Los Angeles. Cultural differences, budget limitations and bureaucratic constraints have caused the process to morph. The COMPSTAT model as employed in Los Angeles, although different from its East Coast cousin, has also led to dramatic crime reductions over the last five years. Indeed, an inherent strength of COMPSTAT and PerformanceManagement is that they can be modified to direct and control significantly different environments.

When I arrived in Los Angeles, the LAPD, like the NYPD in 1993, was focused on corruption control and although the police officers here were arguably among the best trained anywhere in the United States, they were handcuffed by an oppressive bureaucratic structure that stifled creativity and held all of the power and responsibility in a highly secretive and centralized management structure. Fear ruled the troops and commanders at the area level (the LAPD is broken down into 19 Area Commands that in turn report to four geographic Bureaus) were held responsible for what occurred, but had little discretion to affect their own fate, since decisions were made at the bureau level or higher. Crime was up and morale was down. The relationship with the community was deplorable and line officers felt as though their only mission was to answer assigned radio calls and to stay clear of trouble. It was like a dysfunctional family where everyone knows something is wrong, but they are powerless to change the situation. Crime had risen for the three years prior to my appointment as Chief. The officers appeared tired and burned out from what was described by many as a relentless, vindictive and unfair discipline system. The Department was also suffering from lowered and poorly defined goals and expectations.

A major Performance Management change over the last five years is a management philosophy that encourages and rewards risk taking and innovation. Vision, goals and expectations are clearly defined from the top, while line managers are given more freedom and resources to manage their commands in exchange for results responsive to Department goals, as well as enhanced accountability; the accountability brings a capacity to explain what was done and why it succeeded or failed.

COMPSTAT and the LAPD COMPSTAT Plus

By immediately embracing COMPSTAT when I became Chief in 2003, theLAPDreduced serious crime by 4% in the first year, with homicides alone dropping by 21%. But as my Assistant Chief of Operations at the time, George Gascon, recognized the LAPD version of COMPSTAT was viewed as too confrontational and was not the most effective way to assist the most under-performing commands suffering from the most complex problems. For these commands, Gascon theorized, a more in depth analysis would be required to augment the traditional COMPSTAT process, and additional performance enhancers would be in order. Consequently, Gascon and his team created and implemented COMPSTAT Plus.

COMPSTAT Plus represents an enhanced application of the well-known COMPSTAT principles of inspection and accountability, as well as the use of more in-depth auditing methods, mentorship, and close collaboration. To proceed with the inspection, we assembled an inspection team composed of proven experts in the fields that were the subject of COMPSTAT Plus’ focus (patrol and detective operations, the crime analysis section, communityrelated crime prevention efforts, and management and supervision). The group developed a set of inspection protocols to help uncover performance inhibitors with a clear focus on helping the area in reducing Part I Crimes. This goal would be achieved by implementing procedural efficiencies and creating an all-encompassing crime-fighting blue print designed to bring the various stakeholders together with a clear mission.

COMPSTAT Plus’ effectiveness is based on three clear strategies.

Based on an emphasis on targeted performance management strategies and a new and improved version of COMPSTAT, Los Angeles is now the nation’s second-safest large city. In the last year, serious crime has fallen to its lowest point since 1956, while arrests have steadily increased. In a city that saw 2,000 homicides in 1992, the number of homicides in 2007 fell to 392. Still too many, but a vast improvement considering the population of Los Angeles has grown significantly in the intervening 15 years.

Looking forward---COMPSTAT and its implications

As I noted earlier, each application of COMPSTAT must fit the situation and there are a couple pitfalls and misconceptions that we should clear up before proceeding. First of all, COMPSTAT is not all about the numbers. It is about accountability.As I continue to emphasize with my management team, numbers do not tell the whole story and their success in the organization is based on their ability to demonstrate what they are doing to fight crime and to meet our organizational expectations. I would caution police leaders about two commonly encountered problems, as agencies move to aggressively implement an out-of-the-box or off-the-shelf COMPSTAT performance management system. The first is what I call the ‘process trap’ and the second is the ‘gotcha syndrome’. These are two different but connected phenomena. As far as process problems, if your subordinate commands think it is all about preparation for the next COMPSTAT meeting, they are missing the point and will fail. The stress caused by a Captain’s appearance at COMPSTAT can either drive innovation or it can become paralyzing in and of itself. This is where mentoring and coaching among command staff can benefit the entire organization.

What I mean by ‘avoiding the gotcha syndrome’ is that COMPSTAT should be setting your leaders up to succeed. That may sound funny to some of you who may have attended or read about COMPSTAT sessions presided over by Jack Maple and Lou Anemone in New York or by George Gascon in LA, but this process is about success---measuring success, inspiring success, turning underperformance into achievement. Management by inquiry, as the system has been described, is most effective when it is based on authentic and honest dialog. Unfortunately, ‘COMPSTAT-like performance measurement systems can devolve into a forum for negative reinforcement, whereby field commanders can be subjected to embarrassing or harassing questioning’ (Bratton and Smith, 2001). COMPSTAT inspections are most effective when conducted in a collaborative atmosphere that encourages discourse and respect for participants while avoiding prejudgment and heavy handed questioning.

Through a form of participative performance management, the LAPD has been able to achieve consistent year-to-year crime reductions for the past six years. However, this consistency has not been reflected nationally. In fact, the Police Executive Research Forum (2006) and others have noted that there is a ‘gathering storm of crime’ on the horizon. Many U.S. cities continue to register crime increases. In order to sustain reductions and in some cases to further reduce crime and the fear of crime, we as police leadersmust continue to seek new ways to more quickly measure and improve performance. We can look to the private sector, where performance is measured ultimately by the bottom line of profit and loss or successful community initiatives. Although our bottom line is reduction of crime rather than production of profit, business models of performance metrics are illustrative of just how detailed we can get in measuring the efficiency and effectiveness of our senior managers.

Large retailers measure performance in real time according to a series of integrated performance metrics that measure and then display weighted data on everything from comparable gross sales among stores to per square foot sales and even the percentage of customers who enter a store versus those who make purchases. Store locations can be ranked by individual measures and by aggregate measures and viewed via a web-based dashboard to assist decision makers in making deployment decisions for products and for store personnel. Target corporation’s team success metrics are driving change within their retail store teams by providing immediate and ongoing feedback to team members on their relative success versus other stores and versus their own previous success. The system is automated to the degree that very little store personnel time is spent on gathering and analyzing data, allowing them to focus on achieving their individual performance goals.

The future for police department performance management lies in our ability to learn from each other and from the private sector in setting meaningful goals for personnel throughout the organization, implementing a comprehensive set of measures for our precinct level managers and then holding them accountable through a process of ongoing examination of progress against organizational goals. The combination of inclusion in goal setting and empowerment in decision making to achieve goals, along with a process of ongoing feedback can have dramatic and immediate results on the bottom line. In fact, a number of studies have shown that goal setting, in combination with participation in decision making and objective feedback, yields significant productivity gains. These gains occur when top management commitment and participation are high according to these studies (Rodgers and Hunter, 1992).

Cultural change has occurred here at the LAPD and we now have institutionalized leadership practices that are fully invested in performance management. We are well positioned to drive that change to the next level by continually refining our analysis processes and improving our ability to provide managers with real-time feedback on their progress against agreed-upon goals and established benchmarks. From my perspective as a police leader, I also view the future as an opportunity to expand our partnerships and collaborative efforts with our academic and business communities in order to continually improve our ability to forecast crime and to measure our performance in preventing and responding to crime trends.

From a business analytic’s perspective, experts foresee the expanded use of streaming data analysis, pre-programmed threshold alerts and improvements in the way in which we visualize the data analysis that the computer performs on its own. We will move from near real-time analysis to true real-time analysis and then to a ‘predictive policing’ posture wherein more accurate and reliable probability modeling will be utilized to forecast potential crime trends over an increasing time span.

Data analysis

American Police Chiefs used to look at crime numbers annually, in response to the FBI’s Uniform Crime Reports data collection efforts. Beginning in the 1990s we began to critically analyze crime data on a monthly, then weekly, then daily basis. When I arrived here in LA, we developed methods to draw down data several times per day for analysis. The problem we continue to confront is both the accuracy and the timeliness of the data. We must develop ways to shrink down the real-time continuum both on the input side of the equation and on the analysis side of the equation. Is data truly real time, when it may take days for a crime report or field interview card to be entered into the system and available for analysis? Likewise, is data analysis truly useful, when the output is so difficult to interpret that only a handful of analysts at the precinct level can interpret it for use in the field?

We need to move to a model where the computer is constantly querying and displaying easy to understand data displayed in a user friendly format that can be turned into actionable intelligence by our cops on the beat. We are experimenting with ways to provide our patrol units with alerts and data in the field when and where they can act on it, without them having to be trained as crime analysts. Soon, we will see LAPD officers receiving information in relation to their position in time and space via Global Positioning Satellite technology and then acting on and reporting their actions as they are happening for immediate inclusion in the data set. By streamlining data entry and automating it, and then developing a more robust capability to data mine, we will move closer and closer to real time.

Increased use of wireless downloads as vehicles and officers with portable devices pull into the stations will allow us to auto-download data, clean it and make it available for immediate search and analysis. We are researching ways to leverage data to benefit the organization on an enterprise level and then to share that information across organizations on a regional level. Regional COMPSTATs as well as collaborative goal setting and fulfillment among government agencies will become the new standard. An increasing ability to forecast potential performance problems in time to set up appropriate interventions will become the key to success in managing our response to crimeand other problems in society. Computer technology will also likely be used to not only identify possible issues earlier, but to recommend interventions based on artificial intelligence decision support programs and functions that are self-healing and self-correcting.

We also need to partner with businesses at the forefront of performance management to drive our own kind of profitability or performance by measuring everything from crime reduction versus last year or last month to overtime control, sick and injured on duty time usage, morale, community satisfaction, misconduct, excessive force, officer safety, employee wellness and a host of other measures. An appropriate scale and the validity of each measure will need to be sufficiently researched and tested so that we can rely on these indicators in gauging our relative effectiveness. In this context, organizations must move to create an atmosphere and an ethos that encourages managers to make decisions based more on evidence than on intuition. As in the private sector I anticipate and would encourage the establishment of a strategic analysis apparatus within police departments to develop models that will inform those decisions.

Challenged by rising crime rates in the 1970s and 1980s, in the 1990s police leaders rose to the challenge and developed performance management systems like COMPSTAT and success quickly followed. I am confident that our police leaders will continue to thrive creatively in the dynamic environment of the future---an environmentmarked by growing demands on our resources and by an ever-expanding number of specific performance measures. Isn’t that the real measure of a leader, to determine how well we inspire those we lead to stretch and to grow and to develop in response to ever more challenging endeavors? Performance-based management in policing supports what in the end should be every good cop’s bottom line---fewer victims.