Editorial standards: how we source facts, compare competitors and fix mistakes

This page is the working policy behind what we publish on ecomsellertool.com: the blog, the glossary, the comparisons, the AI agent pages and the calculators. It applies to every page we publish or update from September 25, 2026; older pages are brought up to it when we next update them. We are a tech agency that builds software and AI agents for ecommerce brands, so we have a commercial interest in what we write. The rules below exist so that a reader, a journalist, an AI engine or a company we compare can see how a page was made and hold us to it.

How do we choose what to write about?

Topics come from three places: the questions brands ask about running and growing on marketplaces, which we see in search data, in AI assistants' answers and in our own conversations with brands; the changes Amazon and other marketplaces make that affect day-to-day operations; and the systems we have built on Amazon's seller APIs since 2017. A topic earns a page only when we can answer it from primary sources or from our own build work, and when the answer changes what a seller would do next.

Every page has to add something a reader cannot get from the pages that already rank: a comparison table, a formula, a checklist, a real product screen or a note from a build we shipped. If we cannot add one, we do not publish. Before a page is written, we check its target query against every page we already have, so no two of our pages go after the same search and a reader lands on the one page that answers the question.

Google's spam policies (checked September 25, 2026) treat many pages generated mainly to manipulate search rankings, rather than to help people, as scaled content abuse, no matter how those pages are created. We treat that as the floor, not the target. A page that has earned no search impressions and no AI citations after six months is merged into a stronger page or rewritten, rather than left up to pad the site.

Where do our facts come from?

From the source that owns the fact. A fee comes from the marketplace's own fee page, a policy from the policy itself, and a competitor's model from its own website or, for a public company, from its filings with the U.S. Securities and Exchange Commission, which anyone can read on the SEC's EDGAR system. News articles and third-party blogs usually point us to a fact, and we then cite the primary source they point to, not the article. Reporting has a narrower place of its own. When a marketplace has published nothing itself, for example about which vendors it stopped buying from or what it announced on stage at a conference, we may cite a named news outlet, analyst or consultancy, say in the sentence whose reporting it is, and never use it for how a product, fee or policy works. An analyst's own survey or index is cited as that analyst's data.

Each fact is linked inline, on the sentence that uses it, and facts about Amazon, marketplaces, fees and competitors carry the date we checked them. There is no separate list of sources at the bottom of a page, and the structured data behind each page cites only the sources a reader can click on that page. When a feature is in beta or still rolling out, the page says so, for example "beta, US" or "rolling out (US)". When we cannot verify a fact against its source, we leave it out.

Numbers about Ecomsellertool come from one place, our brand facts page: we have built on Amazon's seller APIs since 2017, shipped 50+ tools, served 100+ teams and built integrations for 20+ marketplaces. A number about us that is not on that page does not go into any page written under these standards, however good it would look in a headline. Who we are and how the company started is on About us.

How do we compare competitors fairly?

We start by saying we are not neutral. Ecomsellertool sells a competing service, and every comparison page opens with a note that we are one of the options compared and that facts about other companies come from their own public pages and filings. The FTC's policy statement on comparative advertising (checked September 25, 2026) encourages naming competitors, while requiring clarity and, where needed, disclosure so the comparison does not deceive. Section 43(a) of the Lanham Act, 15 U.S.C. 1125 (checked September 25, 2026), covers advertising that misrepresents another company's goods, services or commercial activities. We write every comparison expecting the company we name to read it.

You can see these rules at work in the Pattern alternatives roundup, where each company's model is described in its own words from its own pages, with the date checked, and a section says when Pattern is the better choice. They also shape accelerator vs aggregator vs agency, which puts five growth models side by side, ours included, on the same rows, and says when each of the others is the better choice.

We hold our own model to the same standard. We never buy a brand's inventory, the brand stays the seller, and on every page written under these standards ownership is described with one sentence: "You keep your accounts, your data and the custom code we build; the Growth OS base is licensed to you." The base system is described on the Growth OS page.

How is AI used, and who reviews each page?

We build AI into ecommerce operations, and we use it in our own writing the way we ask brands to use it: for the routine, high-volume work, with a person making the calls that matter. AI tools help us gather candidate sources, outline pages, write first drafts and run an independent check of each fact against the source it links to. They save hours of reading and matching. A claim that cannot be matched to its source is corrected or cut before the page goes live, and the check is written to our fact-check log: the claim, the source URL, the date, whether it was confirmed, corrected or removed, and who or what did the check.

Google's guidance on creating helpful, reliable, people-first content (checked September 25, 2026) encourages publishers to make clear who created a page, how automation or AI was used and why, so here it is plainly. Pages are credited to the Ecomsellertool editorial team because no single author writes them. Our policy is that a named person, our founder or a team member the founder names, reviews each page and its fact-check record and is recorded in the log with the date. We are not there yet: the pages published under these standards on September 25, 2026 were checked claim by claim by an AI fact-check that did not write them, and no named person has approved them yet. The log says that for every one of those checks, and we will add each human approval to it as it happens. We keep the log with the site's source files and share a page's entries on request.

Some things are never produced by AI on pages written under these standards: reviews, testimonials, quotes attributed to a person, and product screens. Every product screen is a screenshot of software we built, captioned as one account's screen, and it never shows a result as typical. Illustrations, such as blog header art, may be made with image tools. They illustrate a topic and never stand in for proof of a result.

Who reviews each type of page, and how often is it re-checked?

Each page type has its own checks, its own approver and its own re-check cadence, because a competitor's filing, an Amazon rollout and a calculator formula go out of date in different ways. The table is the policy the reviewer works from; the approver column describes the human sign-off we are putting in place, which the pages published on September 25, 2026 do not have yet. Every check it lists is recorded in the fact-check log: the claim, the source URL, the date it was checked, whether the claim was confirmed, corrected or removed, and who or what checked it.

Re-checking means opening the linked source again, not rereading our own page. When a source has changed, we update the page and its date. When a source has disappeared, we find the current one or remove the claim. Each re-check is logged in the same fact-check log as the original review, with who or what did the check and the date, so anyone on the team can see when a fact was last confirmed and how.

What will we never claim?

Some claims are off the table on every page written under these standards, whoever drafts it and whatever the topic:

The test behind the list is simple: if we could not show a reader the evidence for a claim on request, the claim does not go on the page. When a reviewer is unsure whether a sentence passes that test, the sentence comes out, and the page ships without it rather than waiting for a better number. That is also why calculators show an estimate from the inputs you enter and never a forecast of the revenue you will make.

How do we collect reviews and testimonials?

Our review requests and testimonials follow the FTC's Trade Regulation Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465. The FTC's questions and answers on the rule (checked September 25, 2026) say it went into effect on October 21, 2024. The rule text (checked September 25, 2026) covers fake reviews, buying positive or negative reviews, insider reviews and review suppression, among other practices. The rule allows some things we choose not to do, such as disclosed reviews from insiders, so our own rules go further:

Testimonials also follow the FTC's Endorsement Guides. Under 16 CFR 255.2 (checked September 25, 2026), an ad that relates a customer's experience with a central or key attribute of a product is likely to be read as showing what customers will generally achieve, so we never place a client's result next to advice where it would read as the typical outcome. When we recommend a product we are connected to, because we built it, are related to its maker or earn a fee from it, the page says so: 16 CFR 255.5 (checked September 25, 2026) requires an endorser's material connection to the seller, when the audience would not expect it, to be disclosed clearly and conspicuously.

How do you report a mistake?

Email hello@ecomsellertool.com with the page address, the sentence you think is wrong and, if you have one, a link to the source that shows the correct fact. The same address appears at the foot of our blog posts, glossary entries and comparisons. We read every report, including those from companies we compare, and a competitor's correction is handled like anyone else's: if its own current page or filing says something different from ours, its page wins.

What happens after you write:

1. We reply as soon as we can to confirm we have your report and that a reviewer is checking it.
2. The reviewer opens the source again. If the report is right, we correct the page and change the page's updated date.
3. When a correction changes a fact, a number or a statement about another company, we add a dated note at the end of the page saying what changed. Typos and broken links are fixed without a note.
4. If we disagree, we reply with the source we relied on, and the page stays as it is until one of us finds a better source.

If you are here because you are weighing us against another way to grow, the most direct way to test our claims is to see our work on your own data. Start with the free 24-hour diagnostic: connect Amazon with Login with Amazon, no password shared, and get a report within 24 hours of connecting, on business days. Or schedule a call and ask us about anything on this page.