A new study using identical stool samples found that direct-to-consumer microbiome tests can produce dramatically different results—sometimes as different as comparing two separate patients. Despite using similar sequencing technologies, companies varied widely in how they processed samples, analyzed data, and interpreted results, leading to major discrepancies in microbial composition and reported health insights.
What’s striking is that methodological differences—not biology—were driving much of the variation. In some cases, the same sample was labeled both “healthy” and “unhealthy,” with conflicting dietary recommendations. Even clinically relevant organisms showed inconsistent detection across platforms.
This points to a deeper issue: the science of the microbiome may be advancing rapidly, but the infrastructure around it—standards, validation, and clinical interpretation—is still catching up.