Predictive Policing – safer communities or prejudiced aversion?


By: Nikola Rujkov, Mythili Manirajah, Finn Klebe, and Min Wu

Short summary of the service provided by Predpol

Imagine policemen are on patrol right where the crime is happening. Welcome to the world of predicting crime, of predictive policing, welcome to “PredPol”.

The product was introduced in 2012, based in Santa Cruz, California. “PredPol” is a secure, cloud-based software as a service, developed by a team of PhD mathematicians, criminologists, and social scientists. It is an invaluable added tool that allows our police to use their patrol time more efficiently and helps to stop crime before it happens – this technology offers an excellent crime-fighting solution that will ultimately make our lives safer. The core technology includes prediction of drug crime, gang crime, anti-social behavior, and the recently released gun violence prediction tool. In other words, “PredPol” tracks various types of crime: burglary, robbery, vehicle theft, theft from vehicles and gun crime as well as traffic accidents. Only three criteria of data are used to make predictions of the type, place, and time of crimes. No personal data is utilized in making these predictions. This predictive policing gives officers a significantly better idea of when and where to be so that they can deter crime. There is a proven track record of crime reduction in communities that have deployed “PredPol”.

PredPol’s server gathering big data


Key Partners, Key Resources, and Key Activities

PredPol was derived as a project between the Los Angeles Police and UCLA. From the early beginning till nowadays the company is dependent on the information of the previous past types of crimes of the officers and the crime analysts. Because of that, the key supplier of the company is mainly the police. There are three main key resources we are acquiring from our partners: crime type (what happened), crime location (where it happened) and crime date and time (when it happened). While making the analysis, the data is not tied to some demographic groups. Since the crime is based on civilian reports, it requires an actual citizen (or victim, in many unfortunate cases) to report a crime and call for service in their area.

Police departments pay around $10,000 to $170,000 a year to gain access to the company’s information, having heard that other departments that do so have seen double-digit drops in crime. That only amount of money that is given from the police departments. Other income is flowing inside of the company from the analysts’ companies and companies dealing with criminology. Using PredPol’s technology, analysts spend less time making crime maps and more time conducting the analysis and intelligence work for which they are trained for.

Statistical result of predictive policing
PredPol in action – segmentation of focus area (500 by 500 feet)


Cost Structure and Revenue Streams

The costs inherent in this business are mainly service costs. PredPol has been developed over the course of several years both in terms of the mathematics and the software. The team members were mainly in the United States, but the United Kingdom also contributed with some researchers. These costs are mentioned first since they make up the biggest part of the costs. The implementation and maintenance are relatively cost effective. For maintenance, the costs are mainly due to server space, since is it a cloud-based technology. The only men power required to maintain the software would be the local adaptation of the product and it has to be set up differently for every new customer (police departments). The algorithm can stay, but the part of the country where the service is used has to be divided into different squares which are necessary for the algorithm to function and being able to locate the crime acts. However, setting up the software is also done in cooperation with the locals since they are the users. As a result, the cost for the software itself is only paid for once, the maintenance from then on is done by the locals and PredPol starts to extract license fees from their buyers. The license fees are levied since the buyers are granted to use PredPol’s knowledge and patented service for a limited period of time.

The end of the section above, leads us to the section of the revenue streams and the question of how PredPol actually makes money? First of all, before selling the products to their customers, PredPol needs to know what’s the exact value the consumers are willing to pay for. In the case of PredPol, this is the actual service in form of the software which works in different local regions, as well as the fact that officers on patrol will be used more efficiently. This is because of their abilities which are complementary with the software, thus facilitating to solve crime cases, as well as the amount of officers that is sometimes “wasted“ in places where nothing takes place but would have been needed more pressingly in areas with a higher crime rate.

The cost for the consumers, when paying the license fee comprised of service and software fees, depends on the size of the area in which the software is going to be used. To name an example, the city of Modesto, California with 205,000 inhabitants, needs to pay an annual fee of 25,000 dollars, as stated in an article. For Bremen which has around 540,000 inhabitants, the annual fee for the services of PredPol would be $54,878. PredPol especially makes advertisements for the price when saying that the fees are less than police officer’s salaries for one year. However, it is not about replacing, but rather about supporting the police officers, according to PredPol.

Customer relationships and Customer segments

Testimony substantiating the reliability of PredPol

PredPol caters to specific customers and has established a strong customer relationship that has been vital to the success of their service (predictive policing). It has a niche market because it’s service is aimed to fulfill the specific needs of a particular market. It specifically targets the law enforcement agencies mostly police departments and enables them to take preemptive measures and reduce the crime rate in regions that are prone to repeated attacks whilst ensure public safety. For instance, the Alhambra CA Police Department reported a decrease in burglaries and vehicle thefts in 2013 whereas assaults, burglaries and robberies went down by 9, 11 and 27 % respectively within 2011 to 2012 according to the Santa Cruz CA Police Department. One of the many advantages of having a niche market is that it’s cost-effective. To be specific, Predpol spends comparatively less on marketing to reach out to its customer segment. By doing so, it could add more effort and spend extra money to refine its service and subsequently produce one that is of better quality which in turn strengthens customer relationship. However, a niche market could also put Predpol at a disadvantage. For instance, it can’t expect a large turnover since it only focuses on a small group or experience business growth (it is only likely to occur pace unless it expands the service to others).

Predpol connects with its clients on a personal level by providing a more personalized service. To be specific, it tailors its service to predict the possibilities of crime for specific regions instead of providing more generalized results. Predpol’s commitment towards its personalized service is evident when examining the way in which it generates the predictions. It takes advantage of the pre-existing crime records (available in a particular region of concern) that are safely transferred and analyzed to produce accurate crime predictions. In a way, it reproduces it for a different purpose at a minimal cost as the company only expends money to maintain the software. In addition, PredPol also offers a cost-effective service to its customers. To be specific, it helps to reduce the municipal costs because reductions in crime from predictive policing save law enforcement, courts, jails and communities hundreds of thousands to millions of dollars per year. It also allows the police officers more free time to engage with the community on a more personal level and build a strong relationship. Overall, it is evident that PredPol has a very strong foundation in customer relationships and segments that have been crucial to its development and success.

PredPol in action – first successes depicted in statistics


Value proposition and Channel

The purpose of predictive policing is to give information on identifying the potential criminal activity so that the officers can reduce or stop the crimes from happening. However, the traditional way of hot spotting by human analysts often gives us large-scale area which is neither effective nor tactical. We need a tool that gives us great accuracy and accessibility to effectively detect crime due to its time sensitive nature and PredPol does just that!

Predpol uses an algorithm based on the patterns of previous criminal behaviors and then translates the three key facts – type, location (in areas as small as 500 by 500 feet) and time of the crime – into customized predictions for individual law enforcement agencies, while personal identifiable information are not included. The software runs continuously to upload the crime report information to PredPol server so that the customers are always up to date. As for accessibility, the crucial step is to get the information to the officers in the field who can actually act on it. The information from PredPol can be accessed on portable computers, tablets, and even smartphones. The customers can actually see not only the predictive analysis but also notes on what the other officers have witnessed in the target area and a list of the latest crime.

In this way, the law enforcement agencies can plan in advance to deploy resources strategically which in turn increases the effectiveness of policing efforts dramatically. When predictive policing is fully in place, the technology can enable law enforcement agencies to reduce crime as much as 20-30 percent. The benefits of such significant crime reduction could mean less violence, fewer victims, reduced costs, the enormous social and economic benefits, maximized citizens’ security, and a stronger and more stable community.

Since PredPol is a tool for predictive policing, it has targeted and limited customers. What makes PredPol stand out from the other predictive policing tools is the idea of “More Than A Hotspot Tool”. The marketing managers from PredPol try to sell its products to the police officials with this advertising slogan. PredPol claims that it’s different from the traditional approaches of crime maps since its algorithm uses the originated in earthquake prediction. Crime and earthquake share the similarity in the way that although it might be hard to accurately predict the first occurrence, it is still possible to predict the following repetitions. On the marketing side, the metaphor of the earthquake has a great advantage, since in a way it sends customers a message stating that PredPol has already figured out that crime is compatible to mathematics, and this is exactly the solution to the problem that has hindered criminal analysis for years. This is how PredPol uses “real science” to the fight against crime and how it tries to stand out from other similar companies.

Videos about PredPol: 

Orange County Sheriff’s Office uses powerful new law enforcement tool

New police motto: To predict and serve?

Since this final project is posted on the blog, we want you – our reader – to be active and think about the issue at hand. Please consider taking some time to answer our discussion questions on the comments.

1.Do you think predictive policing is a morally defensible tool in criminology or a violation of the presumption of innocence?

2. Should this tool be accessible commercially for private business, e.g. when taking into consideration where to move to ((un)safe neighbourhood)?


Works Cited:



4 thoughts on “Predictive Policing – safer communities or prejudiced aversion?

  1. For question 1, I was thinking about just this issue when I was reading your article. I think there has to be a balance between respecting people’s ability to change (e.g. not all people who commit a crime will be repeat offenders) and preventing further damage. Ultimately, nobody can predict the future, especially with unpredictable matters such as human behavior. Just think about how many times the autocorrect on our phones creates funny situations by misinterpreting what we mean, and that’s much easier to predict than a crime. On the other hand, I do see the advantages of the method, I just think we should be cautious.


  2. Thanks for your response. Ultimately, it depends on the user and how he or she uses the service provided by PredPol. It could go either way as it is entirely dependent on individual perceptions but so far it has been beneficial to the public. We have to wait and see of anything changes….


    1. Also I am not sure about the actual effectiveness of PredPol compared to other possible systems. As a RAND study concluded “Simple heuristics have been found to
      be nearly as good as analytic software in performing some tasks.”. Unfortunately, this is very herd to test, especially as the official crime data (you might want to use to analyze the efficiency) would be subject to the same bias that has already been in the data and that is leading the predictions (in other words you can’t know whether you miss more criminals, which you just don’t notice). So I thing we should demand a rigorous proof of the reliability, before we deploy the systems.


  3. Thanks for your response. Ultimately, it depends on the user and how he or she uses the service provided by PredPol. It could go either way as it is entirely dependent on individual perceptions but so far it has been beneficial to the public. We have to wait and see if anything changes….


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s