TrustList rankings now come in editions, with a reason on every place

TrustList rankings now come in editions, with a reason on every place

On 5 October 2026 we published the first edition of the TrustList rankings, the Q4 2026 edition. It covers 13,736 ranking pages, from accountants in London to vulnerability scanning tools. On 6,519 of them, every number on the page can now be explained. A ranked business carries a short line saying why it sits where it does. A business we cannot yet rank stays on the page without a number, and the page tells you so.

That sounds like a small promise. For us it was a large change. Until this month our ranking pages did what most directory rankings do: they sorted whatever carried the right tag and put a number next to each name. This article explains what we changed, why, and what you will see differently when you open a ranking today.

What a ranking page looked like before

We checked our own page for accountants in London on 5 October, the morning we started this work, and wrote down what we found.

The firm in first place was based in Abu Dhabi. Of the 45 firms on the list, our own rewritten profiles placed about 15 in London. The rest were in London, Ontario, in Winnipeg, Queensland, Milan, Mumbai, Knutsford and Glasgow. They were on the page because each carried the word "london" as a tag, and most of those tags had been scraped from text years ago. Many of the same firms also carried tags such as "us", "me", "co" and "de": fragments of words, not places.

The order was no better. None of the 45 firms had a review, so the old composite score gave them all the same value and the list came out in the order the rows were created. To a reader that looks like alphabetical order with a random first place. The page admitted as much in small print: it said the order reflected profile completeness, not evidence of quality.

And there was no reason on any entry. A firm showed a handful of tags, a "Write a review" button and a number. If you wanted to know why it was third, there was nothing to read.

The category hubs had the same problem at a larger scale. The accountants hub offered 389 sub-rankings. Among them were "Ut", "Dubai," with a trailing comma, a surname listed as if it were a country, and the words "simple" and "variable". Each was a page Google could find and a reader could land on.

We did not need new data to fix this. Our rewrite programme had already produced 57,303 profiles written in our own words, and 53,115 of them end in a block of key facts read from the business's own website: where it is based, when it was founded, how many people it employs, which certifications it holds. What was missing was a ranking that used those facts instead of the tags.

What an edition is

An edition is a dated publication of every ranking that meets our standard. The Q4 2026 edition was built from the evidence we held on 5 October 2026. Its positions stay as published until the next edition in January 2027. If a business improves its evidence in November, it does not jump the queue that day. It appears below the ranked list, marked as having joined since the edition, and takes its place when the next edition is built.

We chose quarterly editions for three reasons.

The first is that a rank should mean something on the day you cite it. A firm that tells a client it is ranked second for accountants in London should be able to point at a page that still says so. Our rank badges now cite the edition: "#2 · Accounting firms in London · Q4 2026 edition".

The second is that a frozen list can be checked. Anyone can compare what we published with what we said our method was, and if we made a mistake, we record a correction on the methodology page with the date, the edition and the pages it affected. We have already used that list. On 6 October we rebuilt the edition twice: once after reading more facts from the profiles we had written, and once after reading software vendors' own security and status pages. Both rebuilds are listed there, with their effect on the count of rankings in the edition.

The third reason is that a quarterly rhythm lets businesses plan. The route up is to publish checkable facts on your own website, earn reviews from accountable reviewers and go through vetting. Those take weeks, not hours, and an edition gives them a date to aim for.

The edition page lists what the edition contains, and the editions index will hold every edition from now on, so movement from one quarter to the next can be compared.

Our standard for a ranking

A ranking page shows rank numbers only when it clears a simple bar. At least five businesses must belong on the page, by evidence. At least three of them must have evidence for a rank number. And every ranked entry must carry a reason.

"Belong on the page" is where most of the work went. A business is on a place ranking, such as London, Manchester or the United Kingdom, only when we can place it there. That means our profile says it is based there or has an office there, or it has an address there. A tag is no longer enough. On 6 October our place data held 141,698 memberships of this kind across 4,299 place terms. A town is given a country from the evidence of its own members, which is how "London" keeps London, Ontario out. The second article in this series goes into place evidence in detail.

"Evidence for a rank number" means at least one fact we can check. That might be a client review from an accountable reviewer, vetting, an award, an expert review, or a fact the business states on its own website that we have read and dated. Founding year, years in business, team size, certifications such as ISO 9001, ISO 27001 or Cyber Essentials, and partner status with major platforms all qualify. On 6 October, 20,165 companies and 1,735 software products had evidence of this kind.

A page that falls short of the bar is still published. Its businesses are listed, but without numbers, and the page says that it is not in the edition because it does not yet have enough evidenced listings. We would rather show an honest list than an invented order.

What you see on a ranking page now

Take the same page for accountants in London. In the Q4 2026 edition it lists 58 firms with evidence of a London base or office, and 17 of them carry a rank number. First is Dixon Wilson: based in London, founded in 1888, an independent firm. Second is Arnold Hill & Co, based in London and founded in 1901. Joint third are Alexander & Co, based in Manchester and London and founded in 1976, and Breckman & Company, based in London with an office in Brighton and in business for more than 50 years.

Each of those lines comes from the firm's own website, read on a date that the page shows. When a firm says "more than 50 years" rather than a founding year, we print "In business 50+ years" rather than inventing a year. When two firms are equal on every measure, they share a place, shown as "=3". The order inside a shared place turns over daily, so that neither is always printed first.

Below the ranked firms come the businesses we list but cannot yet rank. Their entries say what is missing: usually client reviews and vetting, sometimes any checkable fact at all. That is also the route up for any of them, and the page links to it.

Two other changes are visible at the top of the page. A banner names the edition, the date of the evidence and the date of the next edition. A switch offers an SME view: a separate order for small and medium-sized businesses, which the third article in this series explains. For the London accountants the SME view lists 57 firms, 17 of them ranked.

Flagship rankings

Fifty-two rankings get more than a list. These are the pages people land on most, chosen from our Search Console data. Each flagship ranking opens with a short introduction to the category written by us, a podium for the top three with their key facts, a comparison table of the top ten, and a section on how to choose. Where a firm's own facts support it, it carries a "best for" label, such as start-ups, landlords or charities. Each label is drawn from what the firm says about its own clients, never from what we guess.

Every flagship page passed a fit check before publication: we read the ranked entries and removed any business that did not do what the page ranks. That check left 248 exclusions across the flagships. Most were businesses filed in the wrong category years ago.

Where software stood, and what changed on 6 October

Software rankings were the weak point of the first build. A business's own website usually states when it was founded and where it is based, and our rewritten profiles capture that. A software product page usually does not state its launch year, so on 5 October only 542 products had evidence for a rank number. That mattered, because software rankings are among the pages people find most often in search.

On 6 October we read the public pages that serious software vendors publish about themselves: security and trust centres, compliance pages, support pages, service level agreements and public status pages. We read 5,528 vendor domains, and in total added facts for 1,145 products. A vendor that links its ISO 27001 certificate from its trust centre now has that certification on its record, with the page it came from. A vendor with a public status page has a measured count of major incidents over the previous 90 days.

The effect on one page shows why this matters. Our ranking for vulnerability scanning tools had five products and one ranked. That was not enough for the edition, although it is the software ranking Google shows most often. It now ranks three: BurpSuite, with no major incidents on its status page in 90 days; Darktrace, with ISO 27001; and SiteLock, launched in 2008. Products with evidence for a rank number rose from 726 to 1,735 that day.

What has not changed

Nothing paid is an input to a ranking, and that has not changed. Sponsored placements still exist on some pages, and they are labelled "Sponsored" wherever they appear. They never carry a rank number.

A business cannot write its own reason. The reason on an entry is built from evidence we hold, and a business that disagrees with a fact can correct it at the source: on its own website, or by claiming its listing and answering our questions. Claims that a business is "number one" or "the best" are never repeated as fact; a claimed profile that says so carries our note beside it.

And the method itself is public. The methodology page sets out what counts, what each kind of evidence is worth, and how ties are broken. If we change the method, the change is dated there too.

What comes next

The January 2027 edition will be the first to show movement. Each ranked entry will be able to show whether it moved up, moved down or is new, against Q4 2026. Between now and then, the businesses below the line on each page have a clear route to a number: publish checkable facts on your own site, collect reviews from clients who sign in with a company email address or LinkedIn, and go through vetting.

We are also working through the 7,217 ranking pages that do not yet meet the standard. About 3,200 businesses, once they carry a single checkable fact, would bring every page with five or more members up to the standard, and we are rewriting those profiles first.

If you run a business we list and you think a fact about you is wrong, tell us. If we got it wrong, we will correct it and record the correction where everyone can see it.

Sources

• TrustList: the Q4 2026 edition, all editions, ranking methodology and corrections, accountants in London, vulnerability scanning tools and all rankings.
• The facts quoted for ranked businesses come from those businesses' own websites, as read by TrustList for their profiles and shown with the read date on each ranking page.
• Figures about TrustList come from our own records, as of 5 and 6 October 2026: the check of the London accountants page and the accountants hub on the morning of 5 October, the edition build of 5 October, and the rebuilds and counts of 6 October (ranking pages, evidenced listings, place memberships, vendor sites read). They are snapshots and will move as the catalogue grows.