Everybody wants to disrupt health care. The way medicine gets delivered to people in the United States has real problems, and improving lives is noble, laudatory, and lucrative. But health care is hard to disrupt. The same high regulatory bars that ensure safety and efficacy also keep dabblers and amateurs at bay. So when the blood diagnostics company Theranos came along with a transformative new technology, it looked like a unicorn—a unicorn of unicorns, the likes of which that hadn’t been seen since Tom Cruise strapped on pointy ears and went to war with a greased-up Tim Curry.

But here’s the thing about unicorns: They’re not real. Theranos billed itself as an alternative to big-needled, big-billed, big pain-in-the-arm standard blood tests. But on Wednesday, a Wall Street Journal article alleged this was all a lab of cards. Rather than using its hyped finger prick technology, the company was doing most of its work on the same boring machines used by stodgy old diagnostics giants Laboratory Corporation and Quest. On Thursday, the Journal dropped a second bomb: According to a company insider, the FDA has asked Theranos to stop using all but one of its finger-prick tests.

If it’s true that Theranos is a mechanical Turk—seemingly magical technology hiding a sketchy, prosaic secret—then the company’s downfall was in its failure to master science. Specifically, the key would have been microfluidics, the study of moving tiny volumes of liquid. And Theranos isn’t the only diagnostics company claiming to have tamed it.

Microfluidics done right would have all kinds of implications for science—handheld sensors, tiny engines. If it worked for blood tests, it’d mean smaller needles and faster results, and by implication, better care. But liquids are difficult to deal with in small volumes. Think of trying to stir a drop of water with a toothpick. You’re more likely to smear it across the table, or break it into smaller drops. Shrink that volume down even further and bigger challenges emerge. At microscopic volumes, surface tension has more and more of an influence. The fluids get stickier and don’t want to flow without some kind of mechanical prodding.

Microfluids also have low Reynolds numbers, which means they are viscous—more like pudding than water. “One of the big challenges has been how you mix things in small volumes at a microscopic scale,” says Eugene Chan, microfluidics expert and head scientist at DNA Medical Institute in Cambridge, Massachusetts. Mixing pure liquids is hard enough, and it gets a lot harder when the liquids contain relatively large things, like blood cells.

Chan has reason to be worried about those challenges. His company has its own finger prick blood test; DMI’s process starts with a finger prick that draws around five to 10 microliters of blood into a tiny vial. That goes into a machine the company calls the rHEALTH, which mixes in tiny test strips—called nanostrips—and reactive chemicals. The machine shines a laser through the sample, and based on the spectral output created by the nanostrips determines what’s going on in the blood. “We use a single drop of blood to do hundreds of different blood tests,” Chan says.

Currently DMI has two versions of rHEALTH. The first is for researchers, and is used for things like translational medicine. The second, a consumer model, is going through regulatory review by the FDA, as well as clinical trials. Last year he and his labmates won a Nokia xPrize, and NASA is interested in using the technology for missions to Mars. “Each year we present our findings to NASA,” says Chan. “They’re very interested in developing a tricorder.”

At peak rhetoric, Theranos was boasting it could run 30 different tests on a single finger prick of blood. But the Journal’s first article said that their tech wasn’t up to the task, and instead the company diluted the blood and ran it through standard industry diagnostic machines. The second article said the FDA outright recommended that the company stop using all but one test, for herpes simplex I—the only one the FDA has cleared for use outside of Theranos labs.

So why herpes and nothing else? The problem is in the prick—er, the finger prick. Small volumes of blood are great for yes-no diagnostic questions: Do I have herpes? How about HIV? Ebola? “The more precision required for test results, the less likely you’re going to be able to do it from a finger poke,” says Cary Gunn, CEO of San Diego-based Genalyte, yet another finger prick company.

Microfluidics fails when you want tight, quantitative values: How low is my Vitamin D? How high is my potassium? How bad is my prostate cancer? On their way to suck your blood, needles and pricks also slurp up a small volume of interstitial fluid—the stuff that makes up a lot of the 70 percent watery part of you. In a normal draw, this fluid gets diluted in four to nine vials of blood. But in the tiny volume a finger prick pulls, that signal-to-noise ratio gets blown up. Even minuscule quantities of interstitial fluid are enough to throw off sensitive assessments.

Genalyte’s machine, called Maverick, runs 128 simultaneous binary diagnostics. Post-poke, Maverick microfluidically swishes the blood sample across a tiny silicon chip. The chip has sensors tuned to detect 128 different molecules—each indicative of a different thing it’s measuring. “It’s sort of like a tuning fork. You hit the right molecule and you get a certain note,” says Gunn. All in 15 minutes while you wait with your doctor.

Or at least that’s the goal. Maverick isn’t yet consumer-ready. Like DMI, Genalyte is still seeking the FDA’s stamp of approval. It too has a research version that biologists use for lab experiments.

So, here’s the big question: What makes Genalyte and DMI different from Theranos?

Mostly, it’s that the companies are trying to show their work. “If I go to the Genalyte web site I can pull up a list of peer-reviewed papers in top tier journals that describe their technology,” says Stephen Master, director of Weill Cornell Medical College’s diagnostic laboratory. “Or, if I want to investigate DMI, I know their technology is being put through some fairly rigorous head to head comparisons with other platforms.” (Chan also has a paper trail, with patents here, peer-reviewed research here and here, and has presented his data at several diagnostic industry meetings.)

By contrast, Theranos is a black box that has touted results rather than process. “The ability of the lab medicine community to police and correct itself depends on that flow of information,” says Master. Instead, Theranos’ research was internal, and rather than submit their work to peer review the company cited their FDA approvals as evidence that the technology worked.

“I don’t really want to be mentioned in the same article as those guys,” says Chan. “But I guess this shows that you have to publish your stuff, let people know what you’re up to, let your peers look at the data.”

The Journal wasn’t the first to cast a wary eye on Theranos—only the most thorough. Skeptics within the scientific community were wary about the company’s caginess regarding its scientific data, and some expressed discomfort that the company seemed to be gaining industry acceptance because of its charismatic founder, $9 billion dollar valuation, and politically influential board of directors. “There’s probably a perception that the lab community has some desire for Theranos to fail,” says Master. And indeed, this is how Theranos has positioned itself against the Journal’s expose. “By and large the people I’ve interacted with have been very interested in Theranos and their technology,” says Master. “The attitude has been that it would be great if it actually works, and we’re actively looking for approaches that will improve patient care through lab medicine.”

Theranos got a lot of traction by tapping into the frustration—both from consumers and the medical community—that diagnostic testing is too painful, too slow, and too expensive. “Their problem is they tried to do it with existing diagnostic instrumentation, instead of innovating new diagnostic instrumentation,” says Gunn. But the company’s bigger problem is its alleged failures stand starkly against all its previous claims. Both Genalyte and DMI could fail, too. The difference is, those companies failures would be part of an open, scientific exploration—rather than headline news.