We’ve all heard the oft-quoted factoid claiming domestic cats kill billions of birds and small animals every year, and unsurprisingly that number is contested and controversial.

One reason skeptics doubt those numbers is because researchers didn’t observe cat behavior and extrapolate the ecological impact — they handed out questionnaires to owners and asked them how often their cats brought dead animals home. To get accurate results, researchers have to be confident people are answering honestly and have reliable memories. It’s really not the best way to do a study.

So a team at North Carolina State University came up with a better way to measure cat predation on wildlife: They’d take hair samples from more than 400 cats, which would reveal how much of their diets consist of cat food versus prey.

Hair analysis can reveal different isotopes, so the team would be able to directly note each cat’s diet by distinguishing between pet food isotopes and those from prey animals. As the team explained:

A common way to understand the composition of animal diets is to collect samples of fur, nails, or blood from an animal and analyze its carbon and nitrogen isotopes. All organic materials contain isotopes of elements that get locked into body tissues, following the basic principle that you are what you eat. For example, the ratios of nitrogen isotopes present in carnivores are dependably distinct from those of plant eaters. Similarly, researchers can distinguish the types of plants that an animal eats by measuring the ratio of carbon isotopes.

It was a good idea, but the study was derailed by an unexpected discovery: No one knows what the heck is in pet food.

Cat food manufacturers fill their products with mystery ingredients, the team found, which means one bag of kibble or one can of wet food doesn’t have the same ingredients as the next, even if they’re the same flavor from the same company.

Additionally, pet food manufacturers can — and do — change what they put in their products without notifying customers or acknowledging the changes on the packaging.

As a result, the research team couldn’t identify which isotopes were from cat food and which ones were from hunted meals.

“We really thought this was going to be an ideal application of the isotope methodology,” said scientist Roland Kays, a co-author of the study. “Usually these studies are complicated by the variety of food a wild animal eats, but here we had the exact pet food people were giving their cats.”

That discovery essentially rendered the study useless for its original purpose, but like all good scientists, the North Carolina State team realized that failure reveals just as much as success, even if it’s not necessarily what you’re looking for.

They published their results in the journal PeerJ, explaining what they’d learned.

“This isn’t what we aimed to study, but it is important in as much as there are hundreds of millions of cats (perhaps more) on Earth,” said Rob Dunn, a professor in NC State’s Department of Applied Ecology and co-author of the study. “The diets of cats, dogs and domestic animals have enormous consequences for global sustainability, cat health and much else. But they are very non-transparent. In short, at the end of this study we are still ignorant about why some cats kill more wildlife than others, and we have also found we are ignorant about something else, the shifting dynamics of ‘Big Pet Food.'”

As veterinarian Shawn Messonnier put it in an editorial for Pet Age, “the pet food industry remains shrouded in mystery about what’s really inside the pet food bag and how it’s created.”

Calling for more transparency in the manufacture and packaging of pet food, Messonnier pointed out ingredients can have a drastic effect on the health of our furry friends.

“For pet parents, a big leap of faith is required of them because unlike fresh human food, you can’t visually verify the ingredients used, their sources, freshness or the safeness of their handling,” he wrote. “Label language can be difficult to discern, too, so people rely mostly on the observations and opinions of friends and family they trust. Inevitably, people hope what goes in the bowl will translate into well-being and happiness for their dog or cat.”