Provenance

What is real,
and what is not.

This demo mixes three kinds of number. Confusing them is how a founder ends up asserting something in a meeting they cannot defend, so they are separated here rather than blurred into a footnote.

Verified

12

Parsed from a primary source or reported by a named publication.

Industry-reported

2

Widely repeated in industry writing, not peer-reviewed. Directionally useful, not evidence.

Illustrative

4

Invented to show what the screen would look like. There is no trading history yet, so there is nothing to measure.

The claims register

259 authorised health claims, 2,082 refused, across 1,115 substances

Verified

GB Nutrition and Health Claims Register, 19 May 2026

Parsed directly from the GOV.UK spreadsheet. The build fails if any rule cites wording the register does not contain.

213 probiotic strain entries, none with an authorised claim

Verified

Same register, aggregated across strain-level rows

Ashwagandha has no authorised claim, but a claim for the root was submitted in 2008 and is still on hold

Verified

Article 13(1) 'on hold' health claims, DHSC, 31 December 2020. Claim 4194, Withania somnifera ROOT

This entry previously read "none was ever submitted", which was wrong. The on-hold list is a separate publication from the register and had not been checked. Rhodiola (2659), saffron (2038), curcumin (4012) and theanine (2005) sit there too.

The ASA actively enforces against unauthorised supplement claims

Verified

CAP Code section 15; ASA rulings

The market

Care/of raised $46m, Bayer took 70% at a $225m valuation, closed June 2024

Verified

TechCrunch, June 2024

Neko Health raised $700m Series C at roughly $7bn, July 2026

Verified

Neko Health press release; TechCrunch

Nourished holds patents on 3D-printed made-to-order supplements, stocked in Selfridges and Boots

Verified

Company press, NutritionInsight

Holland & Barrett runs in-store nutrition advice across ~800 UK stores

Verified

Holland & Barrett "Qualified to Advise" programme

Around 38% of first-time supplement buyers reorder within 90 days

Industry-reported

Supplement ecommerce industry analysis

Widely cited but not peer-reviewed, and we could not substantiate it from primary research. Treat as directional. The documented company failures are the stronger evidence.

Around 60% of supplement subscription cancellations happen in months two and three

Industry-reported

Supplement ecommerce industry analysis

Same caveat. Plausible and consistent with the failures, but not evidence.

Supplement use declines measurably over time

Verified

US Military Dietary Supplement Use Study, Nutrients 2024

Peer-reviewed. Combination product use fell from 44% to 37% over an average 1.3 years. Not an ecommerce churn figure, but it does establish that discontinuation is real and measurable.

Nutrition and doses

Around 1 in 3 UK adults is vitamin D deficient in January to March

Verified

National Diet and Nutrition Survey, 2019–2023

Reference intakes, upper limits and dietary guidance

Verified

NHS vitamins and minerals; SACN/EVM guidance

Doses in the blend are drafted from these and still require sign-off by a registered nutritionist.

Which nutrients have a usable blood marker

Verified

Standard laboratory practice

Of 24 ingredients, three have a marker worth showing a customer. Serum magnesium misses roughly half of deficient people; plasma zinc is confounded by inflammation.

Business figures

Subscriber counts, MRR, retention curves, cost to acquire, partner earnings

Illustrative

Invented for the demo

There is no trading history, so every figure on the dashboard and partner pages is a placeholder showing what the screen would display. None of it is a measurement.

Unit economics: ingredient cost, fulfilment, packaging, margin

Illustrative

Modelled, not quoted

Real figures need actual supplier and fulfilment quotes.

Marker readings on the retest page, review queue cases, order history

Illustrative

Seeded to demonstrate the screens

The product range and prices

Illustrative

Placeholder

Structure modelled on an existing brand for realism. Not the intended range.

The short version

The register data is the strongest thing here and it is not an estimate. It is the government’s own list, parsed, with the build set to fail if a rule ever cites wording that is not on it. The market context is documented. The two retention percentages are industry-reported and we say so. Every business figure is a placeholder, because there is nothing trading yet to measure.

Back to the start