Our commentary
SciShow's video "What Is a Good Gut Microbiome, Anyway?" spends twelve minutes on a question the supplement industry would rather stayed fuzzy, and arrives at an answer with nothing attached to it. Their case, as we read it: you inherit your gut microbes at birth and accumulate the rest by accident, so nobody fully controls theirs; diversity is the closest thing to a marker of "good"; probiotics are the accessible option and the weak one; and fecal microbiota transplant, the option with real results, is repeated enemas under medical supervision. Nothing is for sale, beyond a Zocdoc sponsorship disclosed in-episode that doesn't touch the science.
That's the rare version of this story, and it's why we're pointing at it rather than taking it apart. Three things we'd add.
The probiotic point is stronger than they made it
SciShow's mechanistic reason is a good one: some gut microbes can't yet be grown outside the body, so they can't go into a capsule. Zmora et al. (2018, Cell) supply the other half. Healthy volunteers took an 11-strain probiotic and were sampled endoscopically rather than by stool; the strains survived transit, but mucosal colonisation was "person-, region- and strain-specific," with some people showing outright resistance — and colonisation was "indistinguishable by probiotics presence in stool." Your stool test can't tell you whether the capsule took. Small, invasive study: mechanism, not a verdict on efficacy.
The verdict comes from the American Gastroenterological Association's 2020 clinical practice guideline, which the video doesn't mention. It recommends probiotics only in the context of a clinical trial for C. difficile infection, Crohn's disease, ulcerative colitis, and IBS in children and adults, and suggests against them in children with acute infectious gastroenteritis. Specific strain combinations are suggested for three narrow situations: preventing C. difficile in people on antibiotics, pouchitis, and preterm low-birth-weight infants. So probiotics aren't useless — that list is real medicine. But notice what isn't on it: no recommendation for improving a healthy person's gut microbiome, because that isn't a clinical indication. That gap is the entire product category.
The FMT trial has a detail the video dropped
FMT is clinical medicine and outside what we cover; we're not evaluating it as a therapy, and nothing here is a reason to seek one or, emphatically, to attempt one. But the video leans on a trial, so the trial's own caveats belong in the record.
SciShow describes Moayyedi et al. (2015, Gastroenterology) as a modest success: 38 ulcerative colitis patients, 9 in remission at week 7, 7 of those 9 from a single donor. All accurate. Two things went missing. It was placebo-controlled — 37 patients got water enemas, 2 of whom (5%) reached remission, a 17% risk difference with a confidence interval from 2% to 33%. And it was stopped early for futility by its safety monitoring committee. "Disappointing success rate" isn't the shape of that; a small real effect in a trial the monitors wanted halted is.
The donor-matters finding holds up as far as it goes. Vermeire et al. (2016) found significantly higher bacterial richness in donors whose stool produced successful transplants (p = 0.012) — in 14 refractory IBD patients, uncontrolled, with no significant improvement among the six Crohn's patients. Wilson et al. (2019), the review that gave us "super-donor," adds a caution the video skipped: one donor picked for two target bacterial families enriched only one of them in recipients.
One number we couldn't trace
The video says non-responding donors carry "almost 800 species" and successful ones "more like 950." Neither figure appears in the text of Vermeire 2016, which reports richness as a p-value and a figure, nor in the Wilson review, which cites Vermeire for this point without numbers. Best guess: someone read them off a chart axis. The underlying claim — richer donors, better outcomes — is supported; the round pair of numbers isn't, so we're not repeating it. Same for "up to 2,000 species," which we couldn't source either.
What's actually ours to write about
Food, and it's the lever with the shortest half-life. David et al. (2014, Nature) put ten people through five days of an animal-based diet and five of a plant-based one. The animal arm moved fast — bile acids up, Bilophila wadsworthia expanded, a sulfite reducer the authors tie to IBD through prior mouse work, which is an association plus an animal model, not a demonstrated human harm. The plant arm produced minimal taxonomic change, and everything reverted to baseline two days after the diet ended. Diet moves the microbiome quickly and it snaps back — not "eat this for a better one."
Exercise, likewise. Zhao et al. (2018) sampled twenty amateur half-marathon runners and found no significant change in alpha diversity; forty metabolites shifted and the family Coriobacteriaceae expanded, offered as a candidate biomarker. The authors also suggested fat and energy intake might be the major factors, so diet and exercise aren't separated here — a caveat SciShow states out loud, and most coverage drops.
A good gut microbiome is a research question, not a product category, and the distance between the two is where your money goes.