Case Study 1 — £179 for a PDF: Reading a Microbiome Report Properly

Theo Vasquez is an illustrative composite carried through this book. ⚠️ The report structure, the methodological sources of divergence and the regulatory position are typical of the consumer sector; the specific company and its numbers are constructed. No real product is described.

Setup

He didn't tell me he'd bought it. He brought it to a follow-up appointment about something else and produced it near the end, slightly sheepishly.

"I know what you're going to say."

⚠️ He didn't, actually. Because the first thing worth saying is that buying it was a reasonable thing to do.

Theo has type 2 diabetes risk, a fourteen-year interest in his own health, and a green folder full of contradictory newspaper clippings (Chapter 1). He was offered a test that promised to replace guesswork with measurement.

⚠️ That is precisely what he has spent this entire book trying to do, and it is the reason the product works.

The report, page by page

The question I asked him first

⚠️ **Not "is it valid." **

"Before you read it — what were you hoping it would tell you?"

He thought about it and said:

"That there's something specific wrong. Something I could fix."

⚠️ That is the honest motivation behind most of this sector, and it is worth naming before any methodological critique, because a critique that doesn't address the wish just makes someone feel stupid (Chapter 17 §17.13).

Working through it together

⚠️ Not debunking. Comparing.

The score

"Where does 64 come from? What's a good score, and who decided?"

We looked for a published reference range. There isn't one. ⚠️ The score is a comparison to the company's own dataset, which is not published, using their own weighting, which is proprietary.

Theo, who works in logistics and understands benchmarking:

"So it's a percentile against their own customers."

⚠️ Approximately, yes.

The Akkermansia recommendation

Page 7's chain, laid out:

1. Akkermansia is lower in his sample than in their reference set. ⚠️ Probably true, subject to §27.3b's caveats.

2. Lower Akkermansia has been associated with metabolic dysfunction in observational studies. ⚠️ True.

3. Therefore his low Akkermansia is contributing to metabolic risk. ⚠️ Not established — association, direction unknown.

4. Therefore he should take pomegranate extract to raise it. ⚠️ A supplement recommendation with no trial demonstrating that raising it in humans improves any outcome.

💡 Aha moment. ⚠️ Steps 1 and 2 are real. Steps 3 and 4 are not. And the report presents all four in the same typeface.

This is Chapter 17 §17.1's sixth tell — mechanism named, never dosed — with a laboratory result attached, which makes it far more persuasive.

⚠️ Theo's response: "The first two bits being true is what made me believe the last two." That is exactly how it is designed to work.

⚠️ Page 9, which is the one that mattered

I asked him to find oats on the list.

They were in the REDUCE column. So were apples and onions.

Then Chapter 11's checkpoint, which he still had: ⚠️ his fibre intake was 21 g/day, and oats, apples and onions were three of his four main fermentable sources.

"So if I'd just done what it said…"

⚠️ He would have reduced the one thing in this entire chapter with good evidence behind it (§27.4), in order to raise a genus whose causal role is unestablished, using a supplement with no outcome data.

We didn't need to discuss validity after that.

What we did instead

⚠️ The report cost £179 and was not entirely wasted, because it produced a conversation that wouldn't otherwise have happened.

⚠️ And one thing I hadn't expected:

"Can I ask — is there anything that WOULD tell me something useful?"

Yes. ⚠️ A fibre count, a food record, his existing blood panel (Chapter 26), and — if he had gut symptoms, which he doesn't — a proper assessment.

He wrote that down.

⚠️ Six months later

Fibre 29 g. Kefir most days. ⚠️ No further tests.

And an observation he made unprompted, which is the reason this is a case study:

"The thing is, it did work. Not the way they meant. I've eaten better for six months because I spent a hundred and seventy-nine quid on a piece of paper, and I wouldn't have done it if you'd just told me to eat more beans."

⚠️ That is uncomfortable, honest, and it complicates the ❌ verdict without changing it.

The test did not measure anything actionable. ⚠️ It did function as a very expensive intervention on his motivation — which is Chapter 3's bundled-intervention problem, arriving with a laboratory report attached.

Analysis

1. ⚠️ Buying it was reasonable. A person who has spent years trying to replace guesswork with measurement is exactly who this product is designed for, and treating that as gullibility misreads it.

2. The score is a percentile against an unpublished reference set. ⚠️ He worked that out himself once asked the right question.

3. ⚠️ The four-step chain is the mechanism of persuasion. Steps 1 and 2 true, steps 3 and 4 unestablished, all four in the same typeface. "The first two bits being true is what made me believe the last two."

4. ⚠️ Page 9 recommended reducing three of his four main fermentable fibre sources — the specific, concrete harm, and the thing that ended the argument.

5. He asked the right question about the retest unprompted: ⚠️ "What would I do differently based on the answer?" That question defeats most testing.

6. The declined supplement bundle was £468/year — ⚠️ roughly fifteen years of home-made kefir.

7. ⚠️ And it worked anyway, for the wrong reasons. Which is honest to report and does not make the test valid — it makes it an expensive way to buy motivation, and there are cheaper ones.

Discussion Questions

1. ⚠️ "It did work. Not the way they meant." Does a product that improves behaviour through an invalid mechanism have value? Chapter 3 and Chapter 19 both raised this. Has your answer changed?

2. The four-step chain puts true and unestablished claims in the same typeface. ⚠️ Design a labelling requirement that would fix it. What would the industry say?

3. ⚠️ The report is sold as wellness, not diagnostics, so no regulatory framework applies (Chapter 16 §16.1). Should it? What would regulation cost, and what would be lost?

4. "What would I do differently based on the answer?" ⚠️ Apply that question to three other tests people buy. How many survive?

5. ⚠️ The consultation used comparison rather than debunking. Would a direct "this is pseudoscience" have worked? What would it have cost?

6. ⚠️ He spent £179 and ate better for six months. Would you have told him it was worthless? Is that the same question as whether it was valid?

Your Turn

⚠️ If you have a microbiome report, or are considering one.

Step 1 — Before you buy, or before you read it again, answer honestly:

⚠️ "What am I hoping it will tell me?" _ ⚠️ "What would I do differently based on the answer?" _

⚠️ If the second answer is "eat more fibre and fermented food" — you already have that, and it's free.

Step 2 — If you have a report, find these four things:

⚠️ In my experience the answer is zero out of four.

Step 3 — ⚠️ Run the four-step chain on one recommendation.

1. What did they measure? _ 2. What is that associated with, in published work? ⚠️ 3. Does the report treat the association as causal? _ ⚠️ 4. Is the recommended action supported by a trial in humans?

⚠️ Steps 1 and 2 usually survive. Steps 3 and 4 usually don't. And they are printed identically.

Step 4 — ⚠️ Cross-check the food lists against Chapter 11.

Does the "reduce" column contain any high-fermentable-fibre foods? ____

⚠️ Oats, apples, onions, garlic, leeks, beans, barley. If any appear in a reduce column, that is the finding, and it is the reason not to act on the rest.

Step 5 — And do the audit the test can't:

Fibre: ____ g/day (Chapter 11) · Fermented food: ____ servings/week (§27.9)

⚠️ Those two numbers are worth more than the report and cost nothing to obtain.