Flock license plate readers cost city big, deliver little | Hacker News Reader

Flock license plate readers cost city big, deliver little

sandiegouniontribune.com

"In San Diego police’s February 2026 Annual Surveillance Report, the department discloses that officers conducted more than 244,000 searches of the Flock automated license plate reader technology in 2025. Those searches played a part in “advancing 361 cases.”

That is an outrageous 99.852% rate of ineffective searches.

The police also disclose that the cost of the system will be over $2 million this year — $2,012,500.

A 99.852% ineffective rate means city leaders will spend $2,009,521.50 on license plate reader technology that does not help any case."

This is a flawed methodology for measuring success.

Solving a case isn't a single correct search. It's a tool, and a single case could have hundreds of searches associated with it.

As more regulation comes in, as it should, we should get much better auditing data that link each and every search to a specific case. This is evolving quickly at the moment, but ultimately it's up to the public to begin to push for requirements like these.

Currently departments do not necessarily require a case number, as many times a case number has not been created yet.

I think a more fair method to measure success is look at how effective each dollar spent on LE accounts for the whole picture. How much more effective did ALPR make each officer/detective on the force? Generally speaking, these are force multipliers and are much more effective than spending on pure body count. Many departments cannot fill seats even if they wanted to.

The quote says advancing a case, not ‘solving’ it. So yes, that number is exactly what you’re asking for.

Setting aside the privacy implications (which are obviously very important), it’s like saying “I searched my filesystem and it went through 1,000,000 files. I found the file but it was 99.999999% ineffective” so yes, that’s not a valid metric

Unless they’re saying every failed search is big problem because of the privacy issues I guess

Extending that, we don't know whether this prevented costlier or more time consuming methods of investigation, led to closing of cases by arrest or not pursuing someone found to be innocent, or otherwise helped increase efficiency by not assigning officers to patrol duties around Flock areas.

I'm 'active threat model' level of anti-surveillance, but it's worthless to try to base anything off such a premature and incomplete picture.

> A 99.852% ineffective rate means city leaders will spend $2,009,521.50 on license plate reader technology that does not help any case."

That's not at all what it means. The cost of the system is almost independent of the usage rate of the system. The proper math is that they spent $5,575 per case advanced. Is that a reasonable cost?

Your math doesn't include the hourly wages of the people who do the searches multiple by the time spent on them. Granted, I don't have that info, but I'm guessing it's not peanuts.

It could be truly peanuts from something that happens automatically during entering things into the system anyways to something like multiple times the raw cost if it's something like 10+ minutes of manual work per average search.

If I apply the Purpose of a System is What it Does (POSIWID) heuristic, then the purpose of Flock cameras can not be cost effective law enforcement.

I don’t think this is the right argument, and I say that as someone who wants legislation to prevent such data from being accessible to federal agencies.

Those cases that WERE solved using Flock data could have impacts worth far more than $2 million. For example if one kidnapped child were recovered, what is that worth to the parents? What about to society, who can feel secure about their kids and also actually experience higher safety due to better deterrence? That’s worth more than the few thousand dollars you could say was spent to support that one search of Flock data.

Ultimately, the data is also only as good (and bad) as how it is used. If local jurisdictions use it to solve crime and to actually prosecute criminals (rather than letting them go) then it can be very beneficial. A lot of people are frustrated by criminals not being brought to justice. And obviously, if their privacy is lost, that’s a downside, but it’s not the sole thing to weigh here.

$2 million a year to run 244,000 searches that advanced 361 cases... That's about $5,500 per useful search.

Meanwhile every car that drove past one of those cameras got logged, timestamped, and stored. These things aren't not law enforcement, they're mass surveillance with a badge.

>These things aren't not law enforcement, they're mass surveillance with a badge.

That sounds like ChatGPT.

Also that $2 mil is the cost of the system, and doesn’t include all those man hours spent running unproductive queries.

So Flock helped solve a case nearly every day of the year?

Maybe! We don't know, all we know is that allegedly:

> Those searches played a part in “advancing 361 cases.”

What does "played a part" mean? What does "advancing a case" mean?

Reads like the intentionally obtuse language of an analytics guy making the numbers as big as possible without outright lying.

If you're Flock, and an officer ran a search against a Flock database, for a crime that was later solved, regardless of how it was solved, they will crow that "Flock solved a crime".

They never solved any cases, only provided a warm lead once a day. If they solved many, they would be proud and say N cases solved. In this case N must be an embarrassingly small number since they don't use concrete language. It's like offering 5500$ to anyone that offers any information on any crime.

Reply on news.ycombinator.com