As Detroit weighs renewal, ShotSpotter data raises cost-benefit questions

This article was originally published by Michigan Advance.

Determining the effectiveness of a piece of law enforcement technology like ShotSpotter, the gunshot detection system used by Detroit since 2021, is a complicated question — and the metrics to gauge that can change depending on who you ask. 

For community members, there might be expectations of reduced gunfire-related crime, Daniel Lawrence, a senior research scientist at the CNA Corporation who has extensively researched gunshot detection tech, explained. For police departments, on the other hand, the benefit is largely expected to be seen in its use as an investigative tool. 

With Detroit’s contract for the system up for renewal in June, performance data obtained and analyzed by the Michigan Advance raised concerns about the cost of keeping ShotSpotter in the city, which has paid $7 million for its use over the last approximately three years.

Victims, Witnesses and Shell Casings Found in a Limited Number of ShotSpotter Incidents

According to information shared by SoundThinking Inc., the parent company of ShotSpotter, between 2024 and 2025, the Detroit Police Department responded to a total of 24,225 ShotSpotter-triggered gunfire incidents.

In just over 12% of those incidents, shell casings were recovered based on a ShotSpotter alert. Witnesses were located in just over 2% of incidents. And in less than 1% of the cases was aid rendered to a victim by a first responder.

In an interview in November, Detroit Police Deputy Chief Mark Bliss explained the role that ShotSpotter has played in the department — even including in cases where physical evidence is not present.

“ShotSpotter for us, for what we use it for, has been a game changer for us,” Bliss said. “We use ShotSpotter as an investigative lead. It doesn’t solely rest on itself. It’s an investigative lead, it’s a part of a bigger puzzle.”

“Now, there could be cases where there are no gunshot casings there, maybe people pick them up, maybe they were collected,” he added. 

And when there are casings, an alert can be even more helpful. Bliss explained that, even if no one is at the scene of the crime, the NIBIN, or National Integrated Ballistic Information Network, can tie bullet casings at the scene of one incident to the same gun being used in a different shooting.

The Detroit Police Department did not respond to a request for comment on the specific findings of this article.

ShotSpotter is one of a huge range of investigative tools used by the department, including other surveillance technologies like facial recognition and license plate readers. But it comes at a particularly high price tag, which city council members will have to decide this summer if they want to renew it.

“Any tool that the police department is going to use to investigate shooting offenses is probably going to be a benefit to them. It really comes down to the cost associated with that,” Lawrence said. “I don’t think a lot of the conversations around ShotSpotter would be had if it was $1,000 a year for a department to implement. It’s just something that’s very, very expensive and the outcomes associated with it are questionable because you’re not getting these huge reductions in crime that you would hope for at face value.”

And those concerns were echoed by local advocates, including Ramis Wadood, a staff attorney at the ACLU of Michigan, who said the data he had seen from SoundThinking Inc. was “not only embarrassingly low for such an expensive technology but also don’t actually point to real outcomes in criminal investigations.” 

“Recovery of bullet casings or even arrests don’t necessarily point to charges and convictions. They don’t point to closed cases. That was the pitch from the city, that this will help us close cases, that this will help us turn evidence into convictions. I’m not hearing outcome-based statistics from SoundThinking or from the police department,” Wadood said.

Bliss noted, however, that whether or not an arrest or a conviction is made in any given case cannot come down to one piece of evidence, like just a ShotSpotter alert. 

“There’s so many different pieces to an investigation. ShotSpotter is just one,” he said. “It’s an investigative lead, but it’s gonna depend upon witnesses. It’s gonna depend upon evidence, gonna depend upon forensic outcomes.”

But Wadood also questioned the cost-benefit analysis of such a tool.

“Even if the company or the police department is able to point to a handful of convictions that stem from ShotSpotter alerts, we have to ask ourselves, is the number of convictions that technology led to really worth $7 million over 3 years?” he asked.

DPD Does Not Respond Significantly Faster to ShotSpotter Alerts Than Gunfire-Related 911 Calls

Another argument in favor of ShotSpotter is decreasing the response times for police officers to get to the scene of a shooting — instead of having to collect information, including the location, from a 911 caller, which can take time especially if the person calling is shaken up or emotional, the location data is automatically sent to the department.

Within the DPD system, ShotSpotter alerts are also automatically given the same priority as any other gunfire call — the department gives each 911 call a priority between one and three, with priority one being the most urgent to respond to.

“ShotSpotter is a priority one, because that’s somebody there, somebody fired a gun. So for us, it’s preservation of life,” Bliss said. “We want to get out there as quick as we can, because we don’t know if there’s someone shot.”

But in Detroit, the actual time difference is not particularly significant — and depending on how it’s calculated, is sometimes slower for ShotSpotter alerts, according to data published by the city.

For 911 calls indicating shots fired or a gunshot wound, the average total response time — which includes call intake, dispatch and travel time — is about 12 minutes. For ShotSpotter alerts, the average response time is just over 39 minutes. 

Lawrence noted that there can be significant outliers for a number of reasons that can skew data, and that medians can present a more accurate picture of the situation. Using that metric, standard gunfire-related 911 calls have a median response time of 7.6 minutes, while ShotSpotter alerts have a slightly shorter median of 6.9 minutes.

Lawrence was clear, though, that these comparisons are not necessarily an apples-to-apples comparison — a single gunfire event can trigger both a ShotSpotter alert and a 911 call or even multiple of one type of response, and that can skew response time data and make it hard to compare the two directly.

Even still, that data raised concerns for Wadood.

“It is concerning that, for all the money we’re spending on ShotSpotter, the response times are not that great,” he said. “But even if they were a hair better than other 911 calls for shots fired, again, we have to think about whether saving a half minute is a good way of spending $7 million dollars, or if $7 million dollars could be spent to reduce response times in other ways or to prevent crimes in other ways.”

As for the outlier data that shows particularly high response times in a handful of cases, there’s not just one reason why those might exist. But one reason, Lawrence explained, is that if officers have reason to believe that an alert is a false alarm, it may not take priority over other 911 calls. 

“If it’s just an alert from ShotSpotter in an area that’s not really known for gun violence, they might not have as much credence towards it,” he said.

In some cities, research has shown that a large proportion of ShotSpotter alerts show no evidence of actual gun crime at the scene — in Chicago, the city’s Inspector General found that around 9% of ShotSpotter alerts had any evidence of a gun-related criminal offense. 

Based on that kind of information, especially if a police department is dealing with a high number of calls to respond to, ShotSpotter alerts might not take precedence, Lawrence explained.

Similar data is not publicly available in Detroit, and information on the case status of ShotSpotter incidents requested by the Michigan Advance through the Freedom of Information Act has not been provided, nearly two weeks past the statutory deadline to respond. 

“The city’s failure to share this information raises significant concerns about both transparency and accountability around the use of this technology,” Lauren Bonds, the executive director of the National Police Accountability Project, told the Advance.

Bonds said the public should know the details of how often ShotSpotter leads to an arrest so they can determine its effectiveness and whether they want to continue supporting it. Additionally, she said said the technology increases the odds of what she called “false identifications” — instances where police go to a location based on a ShotSpotter alert and make an arrest based on limited other physical evidence, simply because someone was nearby the alert.

“It’s not a supplement for other kinds of policing work. One of the things that we’ve seen with ShotSpotter is that it’s not the most reliable. And so that means it can’t be a substitute for other investigative work to determine where a shot has come from,” Bonds said. “Most of these technologies are also increasing the opportunities for the police to interact with the public, and some members of the public that are intimidated by police, feel harassed by the police, feel over policed.”