Note: This article provides general information. Blood in stool, fever, persistent diarrhoea, night-time symptoms, severe abdominal pain, vomiting, or unintentional weight loss require medical assessment rather than a commercial microbiome test. Clinical stool tests are not the same as microbiome profiling.
What do these tests actually measure?
Most consumer tests analyse microbial genetic material in a stool sample. 16S rRNA sequencing generally classifies bacteria at broader taxonomic levels, whereas shotgun metagenomics may provide more detailed species and functional information. The methods do not share the same resolution, cost, or error profile.
Reports usually provide relative abundance. A higher percentage does not necessarily mean that the organism's absolute amount increased; the proportion can rise because another group fell. Stool is also a sample of the luminal community and does not represent every microorganism associated with the intestinal lining.
Sampling
Collection, storage time, and temperature can affect results.
DNA extraction
Laboratory methods may recover some microbes better than others.
Sequencing
16S and shotgun methods produce different levels of detail.
Software and database
Classification algorithms change the names and proportions reported.
Technical differences at each stage add to biological variability. A result reflects the method used as well as the person's gut community.
Why is a healthy microbiome not one score?
The microbiome varies with age, geography, culture, diet, medicines, bowel-transit time, recent infection, and many environmental factors. Two healthy people can have different communities, and similar functions may be carried out by different organisms.
Although diversity is associated with health in some settings, higher is not always better. One diversity score cannot explain microbial function, metabolites, host response, or clinical symptoms. A percentile derived from a company's private database is not a universal health reference.
Why can the same sample produce different reports?
A 2026 study sent standardized NIST-developed stool material to seven direct-to-consumer services and found substantial discrepancies both within and between companies. Between-provider variability was on a similar scale to biological variability between different donors.
This does not mean that every laboratory is equally inaccurate. It shows that analytical validity must be demonstrated before clinical interpretation or dietary advice. If a test cannot measure the same sample reproducibly, recommendations derived from it are also weakened.
‘You have dysbiosis’
There is no standardized definition or universal diagnostic threshold.
‘Increase this bacterium’
A single relative abundance rarely provides a reliable treatment target.
‘Foods for you’
Advice often repackages general nutrition principles through a proprietary algorithm.
‘Your supplement match’
The jump from a report to a specific probiotic or supplement may lack outcome validation.
A recommendation can be generally healthy without being clinically validated or personalized by the test.
Can it diagnose disease?
Research has identified microbial signatures associated with inflammatory bowel disease, obesity, diabetes, liver disease, and other conditions. Association is not enough to diagnose disease in an individual. Disease may alter the microbiome, the microbiome may contribute to risk, or both may be shaped by shared factors.
Clinical care may use stool culture or molecular panels for infection, faecal calprotectin for inflammation, or appropriate tests for bleeding. These address defined clinical questions with validated performance and thresholds. A commercial microbiome profile does not replace them.
Can it generate personalized nutrition advice?
Current evidence does not support a stool test reliably prescribing individual foods—for example, broccoli for one person and avoiding tomatoes for another. Algorithms connecting result to advice are often opaque, and may not have been shown to improve health outcomes beyond standard nutrition counselling.
Most people do not need a microbiome test to increase variety in fruit, vegetables, legumes, whole grains, and nuts; avoid unnecessary antibiotics; move regularly; or sleep adequately. If a symptom-directed elimination diet is appropriate, it requires structured clinical assessment and reintroduction—not a report alone.
What should you ask if you still consider testing?
Ask whether the test is diagnostic or a wellness product; which sequencing method it uses; whether analytical validation and reproducibility data exist; who is represented in its reference database; and against which clinical outcome the report was validated. ‘AI powered’ is not an answer to these questions.
Also check how long the sample and sequence data are stored, whether they are shared with third parties, your right to deletion, and whether the company sells supplements based on its own test. Data privacy and conflicts of interest matter alongside scientific uncertainty.
Analytical validity
Does repeating the same sample produce a similar result?
Clinical validity
Is the score reliably associated with a defined disease or outcome?
Clinical utility
Does using the result improve outcomes compared with not testing?
Data governance
How are samples, sequence data, and personal information protected?
Does this mean the field has no future?
No. Microbiome science is a powerful research field for infection, drug response, disease subtypes, and targeted therapeutics. The use of selected faecal microbiota-based therapies for recurrent Clostridioides difficile infection shows that validated clinical translation is possible.
Future progress may come from standardized sampling, stronger quality control, absolute abundance and functional measurements, diverse reference populations, and outcome-focused randomized trials. Acknowledging today's limits does not reject future potential; it draws an honest line between a research promise and clinical reality.
Take-home message
We can measure aspects of the microbiome, but we cannot yet turn most results into a reliable nutrition prescription.
At-home tests may provide an interesting data snapshot, but their established clinical value for routine diagnosis, disease risk, or personalized food selection remains limited. Because results vary by method, no single healthy profile exists, and recommendation algorithms are often unvalidated, decisions should rest on symptoms, clinical assessment, and established nutrition principles.
Scientific sources
- Porcari S, et al. International consensus statement on microbiome testing in clinical practice. The Lancet Gastroenterology & Hepatology, 2025.
- Servetas SL, et al. Evaluating the analytical performance of direct-to-consumer gut microbiome testing services. Communications Biology, 2026.
- Hoffmann DE, et al. The Direct-to-Consumer Microbiome Testing Industry Needs More Regulation. Journal of Law, Medicine & Ethics, 2024.
- Shanahan F, Ghosh TS, O'Toole PW. The Healthy Microbiome—What Is the Definition of a Healthy Gut Microbiome? Gastroenterology, 2021.
- Peery AF, et al. AGA Clinical Practice Guideline on Fecal Microbiota-Based Therapies for Select Gastrointestinal Diseases. Gastroenterology, 2024.
- Rodriguez J, et al. State of the art and the future of microbiome-based biomarkers: a position paper of the European Society of Clinical Microbiology and Infectious Diseases. The Lancet Microbe, 2025.