The day’s random movie was Ace Ventura—a 1994 comedy in which Jim Carrey plays a pet detective. The Miami Dolphins have their mascot, Snowflake the dolphin, stolen before the Super Bowl, and the search takes 86 minutes of screen time. The film cost $15 million, grossed $107.2 million, became Tom Shadyac’s directorial debut, and propelled Carrey straight from the sketch show In Living Color into the first rank of stars.
The funny thing is, the profession in the title exists—minus the talking ass and the rest of the movie miracles. Real pet detectives work. Kat Albrecht, a bloodhound handler and former forensic investigator, has been finding pets with methods used to search for missing people since 1997, and has built an international network of trained technicians along the way. But what caught me wasn’t the romance of the trade. It was the infrastructure beneath it.
At the foundation of this entire industry lies a thing the size of a grain of rice: no battery, no GPS, and inside it—fifteen digits, not a single letter of the owner’s name. And the most interesting thing is that it isn’t what breaks.
First, they measured the scale. In the fall of 2010, researchers from the University of Illinois, together with the ASPCA, called American households on landlines and asked whether they had lost a dog or cat in the previous five years. Fifteen percent had—roughly one family in seven had been short a pet at least once. The authors noted separately that no national measurements had existed before then: the subject lived at the level of stories from veterinarians and shelter workers.
Then comes the arithmetic from which everything else grows. Of the lost animals, 85% returned: 93% of dogs, 75% of cats. Fifty-nine percent of cats came home on their own, without anyone’s help. An ID tag with a name and phone number accounted for fifteen percent of dog recoveries.
And now a number worth keeping beside those: among people who picked up an animal and tried to find its owners through a shelter or a newspaper ad, 38% succeeded—46% for dogs, and three percent for cats.
“85% recovered” and “3% for cats” are both true, and it isn’t a typo. They are simply counting different universes: most pets return without ever entering the shelter system at all—through the yard, a neighbor, the porch. Once an animal reaches a shelter, the odds of being reclaimed drop to 10–30% for dogs and below five percent for cats. Between these worlds stands the machine this is about.
The authors also calculated absolute figures: over five years, around 12.96 million cats are lost in the United States, and 3.24 million of them are never found.
An animal chip is a passive RFID tag. No powered electronics: the scanner emits a field, the tag catches energy from it, uses it to activate its transmitter, and gives up its number. A glass capsule slightly larger than a grain of rice is injected under the skin between the shoulder blades. Its frequency and protocol are defined by ISO 11784 and 11785: 134.2 kHz.
Its contents are fifteen digits. Three specify a country code, if there is a national database issuing numbers, or a manufacturer code; one digit is assigned to the animal category; the remaining eight or nine are a unique number. No phone number, no address—the tag knows nothing about its owner.
Hence the central misunderstanding holding this whole story together: a chip is not a tracking device and not a database. It is a license plate that stays silent until someone decides to read it. And after it is read comes the most fragile part of the chain: scanner → registry → phone call.
The frequencies deserve their own section, because this is where everything broke in the most literal sense. The American market was held by two companies—AVID and Digital Angel, with its HomeAgain chip—both operating at 125 kHz and both holding patents on the technology. The rest of the world, by then, lived at 134.2 kHz under the ISO standard adopted back in 1996.
In April 2004, an eight-month-old pit bull named Hadden went missing in Stafford County, Virginia. He had been chipped at a Banfield clinic—a new ISO-standard chip, 134.2 kHz. Ten days later, Hadden was euthanized at a shelter. The shelter scanner passed over the dog twice and saw nothing; he was held beyond the required period in the hope that an owner would turn up. The owner called half an hour after it was all over. Then the dog was scanned once more—with the device the shelter had received in January and put away in a cabinet. “Chip found,” the scanner reported.
It was the first documented case in the United States in which chip incompatibility killed an animal. CBS ran its own check: in the Denver area, 75% of the shelters contacted by reporters did not have a scanner for Banfield chips. The company suspended its use of ISO tags, AVID sued over the advertising, and in July 2004 four associations—AVMA, AAHA, HSUS, and ASPCA—gathered in Washington as the Coalition for Reuniting Pets and Families and demanded that manufacturers produce one scanner capable of reading everything. The letter went out to eight companies at once, including both sides of the patent dispute around 125 kHz. Iams offered to pay for 30,000 such devices. Digital Angel’s president called the demand “immoral”: come, he said, not with an ultimatum, but with compensation for ten years of investment.
The sides in this dispute differed not only on the numbers, but on the facts themselves. Banfield said its chip was a standard “used in 148 other countries”; AVID’s head replied that ISO was not a global standard, and that the world’s best pet-return system operated precisely on 125 kHz. Both sides offered the same thing as proof—the amount of equipment already installed.
A regulator might have settled it, but it had no lever to pull. Congress instructed the Department of Agriculture to draft rules; in July 2007, USDA-APHIS replied that the Animal Welfare Act gave the agency no authority to regulate private pet ownership. The United States still has no national standard.
The technical side was not a marketing invention, either. In 2008, JAVMA published a direct comparison: chips sold in the United States came in three frequencies—125, 128, and 134.2 kHz—and two protocols, encrypted and open. Four commercial scanners, including three “universal” ones, were tested on dogs and cats in six shelters. Not one showed one-hundred-percent sensitivity, and the result depended on the chip’s frequency, its orientation under the skin, and the number of passes. Hence the practice shelter workers sometimes know better than the manual: move the device slowly across the back, at different angles, and more than once.
In the end, the market solved the problem. Truly universal scanners appeared, and at the end of September 2009 AAHA launched a unified online lookup: enter a chip number and find out which of dozens of private registries holds it. Tellingly, the state was moving in the opposite direction at the same time—Britain simply made identification mandatory: from April 6, 2016, every dog owner had to have a chip or face a fine of up to £500; in 2024, the same rule was introduced for cats in England. They made chipping mandatory and left the registries private.
You would think that if the chip has been scanned, everything is fine.
No. The most detailed study here was conducted by Linda Lord’s team at Ohio State University: 53 shelters in 23 states, 7,704 chipped animals, monthly records from August 2007 through March 2008. Owners were found in 72.7% of cases. At intake, the number was read in 87%; another 10% were found during medical examinations, and 2.5% only on the threshold of euthanasia. Without repeat scanning, 12% of tags would have gone unnoticed.
Then come the reasons for failure—and they are the answer to the entire article: an incorrect or disconnected phone number (35.4%), an owner who does not answer calls and letters (24.3%), a chip registered in a database other than the manufacturer’s (17.2%), a chip not registered at all (9.8%). The number has been read, the technology has done its job, and the chain breaks on the phone book.
“A microchip is only as good as my ability to keep my information in the registry up to date,” Lord put it.
Six years later, Australians tested the same thing on another continent and got harsher figures. RSPCA Queensland shelters received 15,122 animals—7,669 dogs and 7,453 cats; 28% of dogs and 9% of cats had chips. 37% of chipped animals had data problems: 48% of tags were registered to a previous owner or organization—among the latter, breeders, kennels, and rescue groups predominated—29% had phone numbers that were all unanswered or disconnected, and 14% were registered nowhere.
What follows is the arithmetic of odds. When the data were in order, owners were reached in 93% of cases for dogs and 75% for cats. When the data had problems—in 70% and 41%, respectively. Reclaim rates climbed in steps across three groups: for dogs, 87% with clean chip data, 69% with problematic data, and 37% with no chip at all; for cats—61%, 33%, and 5%.
In terms of odds, that is a sevenfold advantage for a chipped dog and a twentyfold advantage for a cat.
Twentyfold—even though cats are the group whose tags are read least often. Cats are less often picked up and taken to shelters: when a found animal is searched for through a shelter or newspaper, owners are located for 46% of dogs and three percent of cats. And yet cats need ID tags most of all—fewer than one in five wear a collar with a phone number.
A separate science is where to search. It had to be built almost from scratch, and one of the co-authors of a 2018 study in the journal Animals was Kat Albrecht from the hook—by then a professional pet detective for twenty years.
They collected 1,210 stories of missing cats. Sixty-one percent were found within a year; a third, 34%, were found alive in the first seven days; after ninety days, virtually no one was found alive. Three quarters of cats turned up within 500 meters of the point where they disappeared. And here is the discrepancy that breaks intuition: for cats with outdoor access, the seventy-fifth percentile of distance traveled was 1,609 meters; for indoor cats that had never been outside, it was 137 meters.
Then comes what Albrecht had really taken up this subject for. An indoor cat that ends up outside does not behave like a runaway dog. It hides—under the nearest shed or in someone else’s basement—and stays silent for weeks. Albrecht called this the “silence factor,” and the related threshold beyond which a cat stops responding to a familiar voice, the “threshold phenomenon.” Hence the practical conclusion, the opposite of the usual one: calling around and posting notices are almost useless for a cat; you have to work on foot, through other people’s sheds, in the very first week. And one more detail from the same study: 46% of cats had a chip, while 19% wore a collar with an ID tag.
Put it all together, and the question from the hook gets a fairly dry answer. The machine works. The tag costs pennies, lasts for decades, needs no power, and does not wear out; scanners have learned to read all frequencies; unified number lookup has closed the gap between registries. When the data are in order, 87% of dogs and 61% of cats return home.
The bottleneck turned out to be the directory: private, fragmented among dozens of companies, and updated manually by a human being. Fifteen digits travel with the animal anywhere and remain inside a tag beneath its skin for however long they like. The owner’s phone number goes nowhere: it is disconnected when they move or change carriers. That junction is exactly where the 35% of calls that reach no one are lost.
In the 1994 film, everything was resolved in 86 minutes and a conversation with an animal. In reality, everything is decided by a line in a private registry that most owners filled out once—on the day of implantation.