The buffer was cut because it never broke.

On a Tuesday morning in mid-April, a beeping sound kept bouncing off the walls of the San Diego County Sheriff’s Office 911 Communications Center. NBC 7 Investigates was in the room for it. The beeps mean at least one 911 caller is on hold, waiting for a person to pick up.

The waiting isn’t new. The California Governor’s Office of Emergency Services sets the standard: answer 90 percent of 911 calls within 15 seconds and 95 percent within 20. The sheriff’s office, the primary law enforcement agency for 910,000 of the county’s 3.3 million residents, has missed that standard four years running. In 2025, 1 in 4 callers waited more than 15 seconds. The average wait was 13.6 seconds. The nine other law enforcement agencies in the county averaged between 1.6 and 5.7 seconds.

The gap is measured, published, and old. This year the county eliminated 13 budgeted dispatcher positions from the communications center.

The sheriff’s office says it reclassified them. Some of the money stayed put and paid for two new communications supervisors and a supervising dispatcher. The rest went elsewhere in the department: six positions to the crime lab, including forensic evidence technicians, and a couple of assistants for human resources. Three more dispatcher jobs were frozen, their funding turned into department surplus. Sixteen dispatcher positions, all told, are out of this year’s mix.

Tamra Samuels, 20 years with the office and one of five coordinators managing dispatchers, gave NBC 7 the rule behind it. “The unfortunate part is our positions are ‘hard-to-recruit positions,’ ” she said. “That’s kind of how they’re defined. So, when positions sit vacant for a period of time, and we cannot fill them, typically they get reallocated elsewhere.”

Read that as an operating policy. A job you can’t fill for long enough becomes a job you no longer have. The vacancy is the evidence used to retire the position, and the shortfall that vacancy created is part of what made the position look expendable.

Nothing collapsed to produce this decision. No single 911 call went so wrong it made the news. Thirteen and a half seconds isn’t a catastrophe; it’s about six seconds worse than the 7.8-second average across the state. Four years of it reads on a spreadsheet as a line that doesn’t move. Cal OES says it notifies agencies that miss the standard, and a county spokesperson said the sheriff’s office has received no formal warning or enforcement action. A Cal OES spokesperson declined to answer inewsource’s questions about whether the office enforces the standard at all.

Your version is smaller and works the same way.

The second truck that sits in the yard most weeks, so you sell it, because it isn’t earning. The Friday afternoon you used to hold open for callbacks and warranty work, which has been booked solid with billable jobs for a year and won’t be given back. The part-timer who answered phones from three to five, whose hours you trimmed because the calls got picked up anyway. The lead time you quote as three weeks that has actually run five since spring.

Each of those has a number attached, and you probably already have the number. Callbacks handled late. Jobs that slipped past the promised date. The share of calls that went to voicemail before somebody answered. The gap has been sitting in your own records long enough to look like the baseline, and a baseline is hard to argue with when you’re deciding what to cut.

What makes a buffer look wasteful is the same thing it exists to do. It absorbs the bad week quietly, so on the spreadsheet it reads as an expense with nothing next to it.

El Cajon Police answers 96.6 percent of its 911 calls within 15 seconds, with an average wait of 5.25 seconds. Its communications manager, Craig Groll, offered no method beyond mandatory overtime for dispatchers and making staffing a priority. San Diego County is working at it too. An AI system called Hyper took over non-emergency calls in March, training-officer bonus pay went up 10 percent, and the answer rate improved to nearly 80 percent over the first four months of 2026. Still under 90.

So the question this one leaves you with: which margin in your operation has slipped below its own standard for so long that the shortfall now reads as normal, and is that the thing you’re tempted to cut because it never actually fails?

← More observations