Methodology: how we research, source and update this site

We label every kind of statement this site makes

We link every policy and legal statement to its primary source

Every source has one entry in a single file in the site's code, with its title, publisher and address. A page cites a source by its key. If a key does not exist, the site does not build. A separate check opens every source address and fails if one no longer resolves.

Each page lists the sources it used at the end. Key facts at the top of guides and service pages carry their own source line.

Trustpilot's help center is rendered in the browser, so we read those pages in a real browser session and keep the text we read. Where a source could not be opened or did not say what a claim needed, we dropped the claim. Two examples are recorded in our notes: a DEA page that refused every request, and an FTC page that did not support the sentence we had drafted.

We quote rules where the wording matters and paraphrase where it does not. Quotations are exact.

We date every page and log what changes

• Updated is the date the page last changed in substance.
• Published appears on guides and is the date the guide first went live.
• Reviewed against Trustpilot's guidelines appears on guides, rule pages and the main service page. It is the date we last checked the page against Trustpilot's current documents.

Trustpilot revises its guidelines and help pages often. When a source changes, we update the pages that cite it and change their dates. We do not change a date without changing the page.

We build comparison pages from public facts only

We read each company's public website: its homepage, its Trustpilot or review removal page, its pricing or cost page, its FAQ and its about page. We record what those pages state and quote them where wording matters.

• When we could not find something, the page says "not found", with the date.
• A company's claims about itself, such as client counts or success rates, are repeated as claims. We say when we have not verified them.
• Links to other vendors carry rel="nofollow".
• Each comparison says what the other company does well and which buyer it fits better than us.
• We wrote these pages and we are one of the two companies on each. Every page says so.

Corrections from a named company are checked against its site and made promptly. The disclosures page covers our conflict of interest in full.

We build the data pages from public profiles

The DTC health Trustpilot report 2026 comes from public Trustpilot profile pages read on one date. We record the TrustScore, the review count, the star breakdown, whether the profile is claimed and any notice Trustpilot shows. A script turns the working sheet into category totals. Only the totals reach this site. No brand name does.

Sampling is not random. We listed well-known US brands in each category. The page states its limits.

We label every example as an example

Review texts, audit results, proof reports and sample audits on this site are written by us to show a method. Each one sits in a frame marked "Example" with a footnote. They contain no real customer data and describe no real client.

We publish no testimonials, client logos, removal counts or ratings of our own, because we will not invent them.

What we never do on this site

• We do not state a success rate for flagged reviews. Trustpilot does not publish one.
• We do not state how long Trustpilot takes to decide a flag. It does not publish that either.
• We do not give legal advice. Pages about law describe published rules and point to counsel.

Related policies

The editorial policy covers who writes and reviews pages. The disclosures page covers how we make money. The policy for AI crawlers covers how AI systems may use this site.