Independent ranking of GEO agencies and (local) GEO specialists

Register for edition 1

Best GEO agency: how to recognise a fair ranking

By Gijs Bodenstaff · published · updated

Most lists of the best GEO agency are written by an agency that names itself, often in first place. A fair ranking states who makes it, what interest they have, how things were measured and how errors are corrected.

Anyone searching for the best GEO agency mainly finds lists by GEO agencies. That is no coincidence and no conspiracy either. It is a well-known pattern, and you quickly learn to recognise it.

In this article we show how often agencies rank themselves, which criteria make a list credible and how we approach it ourselves. Including our own weak spot.

How often an agency puts itself at the top

In August 2026 EchoWi opened 29 readable rankings of GEO agencies in four countries. 21 came from vendors in the category. All 21 named themselves, and 19 put themselves in first place.

The 7 lists from parties with no stake never named themselves (EchoWi, August 2026). These included a CRM company, a tool, trade press, a directory, a podcast and a business school. EchoWi itself also sells GEO consultancy, which does not invalidate the research but is worth mentioning.

It also plays out beyond the lists themselves. In Germany, for "beste geo agentur" (best GEO agency), Google AI often cited a paid, labelled advertorial (EchoWi, August 2026). Why such lists work so well in AI is explained in why you had better not game AI answers.

The Netherlands and Belgium: 23 lists counted

On 30 September 2026 we opened all Dutch and Flemish lists that name GEO agencies. Of the 23 lists, 18 name the maker itself or its parent company, 15 put the maker at the top and 8 disclose that interest with the list. Not a single list came from a party without a commercial interest.

All lists, with link, the maker's position and method, are in our check of 23 GEO rankings. Some makers do it neatly and openly, others do not. A few examples:

  • Gold Lemon ranks 212 providers in its GEO benchmark of September 2026 based on 87 questions in 7 AI engines. Gold Lemon is itself a GEO provider and appears as the "measuring party" in 17th place. Method and weighting are public.
  • Estrelle identified 256 parties in September 2026 and assessed 42 of them in depth. Estrelle, itself a GEO specialist, is in first place.
  • Wildbos included itself in its own list of six in August 2026, and writes that it could not assess the delivery for clients.
  • TEAM LEWIS puts itself in first place in its list of best GEO agencies, without a method.

Pay attention to what such lists do and do not measure. Gold Lemon says so honestly itself: "This index measures visibility in AI answers and a provider's own website, not the results an agency achieves for clients." (September 2026). Being mentioned is therefore not yet proof of good work.

That difference is large. Estrelle found that AI mentions, technical foundation and transparency of offering are barely related (September 2026). At Gold Lemon the most frequently mentioned provider even had an offering transparency score of 0.0 (September 2026).

Also check the claims agencies make about themselves. Estrelle found an unverifiable or incorrect claim on the own site of 4 of the 20 providers (September 2026). One example was a Wikidata registration that did not exist.

Finally, look beyond the lists themselves. In the US, according to EchoWi (August 2026), Reddit was the third organic result for "best geo agencies". Buyers look there for experiences of other clients, and that is a healthy instinct.

Does a self-published list benefit its maker?

Lists work in AI, that much is true. At Ahrefs, 43.8% of ChatGPT's sources were a "best X" list, and 67.6% of 250 search results for "best X software" contained a list in which the writer put itself at number 1 (Ahrefs, December 2025).

But the gain for the maker is small. When AI cites a self-published list, the chance that the maker is recommended rises from about 4% to 7%, and in 24.3% of those answers a competitor is recommended instead (Scrunch, September 2026).

Lily Ray saw in Google AI Overviews that the maker was left out 69% of the time, while its own list served as the source (224 of 323 lists, June 2026). Sites with many such lists lost between 29% and 49% of their visibility in Google in January 2026 (Lily Ray, February 2026).

It plays out locally too. In a local AI answer, Google cited one lawyer's self-promotional list for all the firms in the answer (Search Engine Roundtable, August 2026).

What the law says

A hidden commercial purpose can be a misleading omission (art. 6:193d BW, Dutch Civil Code). Comparative advertising aimed at businesses must be objective and verifiable (art. 6:194a BW), and according to the Nederlandse Reclame Code (Dutch Advertising Code) advertising must be recognisable as advertising.

On rankings, the ACM (2024), the Dutch consumer and competition authority, writes: "If information about the ranking is missing or unclear, it is misleading." We found no ruling specifically about agencies' self-published lists. We are not lawyers; treat this as background, not as legal advice.

What a credible ranking meets

In 2026 the communications agency FL PR listed what a credible comparison discloses. Use it as a checklist:

  • Who the publisher and the author are.
  • What commercial interest the publisher has.
  • Which candidate pool the list comes from.
  • Which criteria there are and how much they weigh.
  • Which evidence threshold an agency has to meet.
  • A reference date and a log of corrections.

FL PR also names warning signs. The publisher is in first place without any calculation. There is no sample, no prompts and no dates. Everyone is praised equally. The year in the title changes without a re-assessment. Or there are undisclosed referral fees.

In January 2026 Fishkin gave similar advice about AI measurement tools. It applies just as much to rankings.

Before you spend a dime tracking AI visibility, make sure your provider answers the questions we've surfaced here and shows their math.

Rand Fishkin, SparkToro, January 2026

In short: do not spend a penny until your provider shows its calculation. A list without a calculation is an opinion, however nice the layout.

Verifying reviews: what Clutch does

Review platforms can help, if they check reviews. According to its own methodology page, Clutch verifies reviews in three steps: logging in, an analysis of the digital footprint and manual editing.

Sponsored spots exist there too. According to Clutch they do not affect the score. That is exactly the kind of disclosure you want to see with every list: what is paid for, and what is not.

How this site approaches it, and where our weak spot lies

Our rankings follow the public LSAI methodology. There is no paid placement. We measure visibility in AI as a percentage over many repetitions, never as a position.

Being mentioned carries little weight with us. Proven client results, reviews and follow-up calls to reference clients weigh more heavily, with the editorial team calling clients chosen by lot. That way we measure what Gold Lemon honestly says it does not measure: what an agency does for clients.

Now our weak spot. Our founder, Gijs Bodenstaff, works in the field himself, as a local SEO and AI specialist. That is why he and his own brands are not ranked. That does not remove the interest completely, and we would rather say so ourselves than have you discover it.

If you spot an error, report it. Every change goes on our public corrections page. The first edition is in preparation; until then there is a provisional list based on public sources, without follow-up client calls.

What you can do now

  1. For every list, first check who publishes it. If the publisher is on it, look for a disclosure of interest and a calculation.
  2. Go through the FL PR checklist above. If criteria, dates or prompts are missing, treat the list as an advertisement.
  3. Ask an agency for client contact details and call a few yourself. Do not pick them from the shortlist the agency hands you.
  4. Put our ten check questions for an agency to every agency you are seriously considering, write down what each of them answers and then compare those answers side by side.

Frequently asked questions

Why is the agency that writes the list often at the top itself?

Because lists of the best agency work well in search engines and AI answers. Whoever publishes such a list wants to benefit from it. Research by EchoWi showed that vendors almost always name themselves.

Is a list that includes its publisher always unreliable?

Not always. If the publisher discloses its interest and makes the method, weighting and data public, you can check the list. Without that openness it is mainly advertising. See also: Measuring AI visibility.

What is the most important warning sign?

The publisher in first place without any calculation. After that: no prompts, no sample and no dates. Then you cannot check anything. See also: AI visibility tally sheet.

Does being mentioned a lot in AI say anything about the quality of an agency?

Not much. Visibility in AI mainly says something about mentions and marketing. Whether an agency does good work for clients is not something you measure that way. See also: Local GEO: how a local business becomes visible in AI.

How do I know whether an agency's reviews are genuine?

Check whether the platform verifies reviews and how. Clutch, for example, describes a check in several steps. In addition, call a few clients yourself. See also: Best (local) GEO specialist.

May a ranking accept money from agencies?

It may, but then it has to be disclosed and must not affect the score. Undisclosed referral fees are a clear warning sign. On this site there is no paid placement. See also: LSAI method.

What is a corrections log and why does it matter?

A public overview of errors that have been corrected, with dates. It shows that the publisher acknowledges mistakes and maintains the list. A year in the title without a re-assessment says the opposite.

How does this site prevent the founder from being favoured?

The founder and his own brands are not ranked. We also state openly that he works in the field himself. Method and corrections are public, so you can check us.

Why do you make follow-up calls to reference clients?

Because visibility and attractive case studies say little about the work itself. Clients know whether an agency delivers what it promises. Choosing by lot prevents the agency from only letting its best client speak.

When will the first scores appear on this site?

The first edition is in preparation. A provisional list based on public sources is already available. The final scores will be accompanied by the method and reference date.

Sources

Further reading in the knowledge base

Book a call