Who Says So? The evidence behind 395 US SEO and app agencies

1. 90.7% of claims check out against their source. A claim is confirmed when the cited page contains the quoted text and supports the claim. 3.3% are only partly supported, 1.2% are not supported by the page, 1.5% point to pages we could not read, and 3.3% are not checked yet.
2. Only 9.5% of claims come from someone other than the agency. Of 10,532 claims, 999 come from an independent source such as an open review platform, a client, or the media. The average agency has 26.7 claims on file, of which 2.5 are independent.
3. 45% of agencies have no independent claim at all. 178 of 395 agencies have nothing on file that the agency did not write or control. 259 (66%) have no independent claim about their actual work.
4. The typical agency's evidence lives on two websites. On average an agency's claims come from 2.2 different sites. For 170 agencies (43%), everything comes from a single site, almost always their own.
5. Names are easy; numbers are rare, and independent numbers are rarer. 365 agencies (92%) name at least one client. 231 (58%) publish at least one measured result: a number, over a time period, recent enough to matter. Of 1,389 measured results, 53 (3.8%) come from an independent source. Only 34 agencies (9%) have a measured result that someone else published.
6. Nearly half of agencies publish no price. 208 of 395 (53%) disclose any price. 170 publish a monthly minimum; the median is $2,500 a month. Among companies that do only mobile app development, 6 of 75 (8%) publish any price.
7. 72 agencies (18%) do not appear in any of our 156 rankings. None of their documented work matches any of the 26 buyer needs we rank. 11 of them have no claims on file at all.
8. Independent proof depends heavily on the need. Across 3,810 ranking positions, 11% stand on an independent source and the rest on the agency's own published case studies and reviews. The share runs from 27.5% for SEO with an in-house team, 26.1% for SEO on a monthly retainer and 20.6% for SEO within $6,000 a month to none at all for Flutter development, International SEO and React Native development.
9. At the top, independent proof wins more often. In 73 of 156 rankings (47%), the first place position rests on an independent source, against 11% across all positions.

A buyer does not choose "a good SEO agency". They choose an agency for an ecommerce site in Miami, a technical audit in Chicago, or a Flutter app in Houston. The useful questions are:

• Has this agency done this kind of work before?
• Who says so: the agency, or someone else?
• Can I open the source and read it?
• What did it cost, and what changed as a result?

Star ratings and review counts answer none of these directly. Our data shows why: for most needs, most of the available evidence is the agency's own account. That is not a reason to distrust agencies. It is a reason to label each claim by who made it, and to rank on that in the open.

That is how Vendar ranks. Each ranking answers one buyer need in one city. An agency enters a ranking only when a case study or client review documents work for that need. Its position is a score computed by the published model; every position lists the claims behind it with their sources, and says whether it rests on an independent source or on the agency's own material. Payment does not change the order. A vendor checklist turns the four questions above into a call script.

Paid agency directories are no longer a source. Directories that sell placement held 439 of the claims we had collected (4%), across 129 agencies. Because a listing there can be bought, we stopped counting those claims as evidence, and they are not part of any figure on this page. For 26 agencies, those claims had been their only independent proof.

Profile verification no longer moves a position. Our freshness penalty used to read the date a profile was last verified, which an agency can pay for. It now reads the last time the claims themselves were checked against their sources, which happens for every agency alike.

• Population. All 395 agency profiles Vendar publishes on 2026-11-02: 256 SEO agencies, 75 mobile app development companies and 64 that offer both. Headquarters: New York 72, Los Angeles 57, Chicago 51, Miami 48, Houston 41, Phoenix 36, elsewhere in the US 90.
• Claims. A claim is one statement of fact about an agency, stored with its source address. Links without a statement and our own research notes are excluded, as are claims hosted on paid agency directories.
• Who says so. Code decides provenance from the source address: the agency's own site and channels it controls are the agency; open review platforms, clients, media and other third parties are independent.
• Check. An AI model reads a saved copy of the source page and decides whether it supports the claim; code then confirms that the quote appears verbatim on the page. Each check stores the model, prompt version and date.
• Work and results. A claim documents work when it is a case study, client review or portfolio project. A result is measured when it carries a number, a time period, and is recent enough.
• Rankings. 26 buyer needs in 6 cities give 156 rankings, computed by the same code that renders the site under model 2026.11.5. 102 of them clear the evidence threshold to be open to search engines.
• Prices. An agency discloses a price when a public source gives a monthly minimum, a minimum project size, an hourly rate or a price range.

• Our coverage is uneven. We researched some cities and needs earlier and deeper than others, Los Angeles first. Part of the gap between cities is our research effort, not the agencies.
• Confirmed is not the same as true. A confirmed claim means the cited page says it. An agency's own case study can be confirmed and still be unverifiable.
• The population is not random. It is the agencies we found and published in six cities, not a sample of all US agencies.
• AI checks can be wrong. Any single result can be reviewed, because each one keeps its model, prompt and date.
• We have a stake. Vendar is itself a way to choose an agency. The per-agency counts behind every figure are published below so the numbers can be checked.

Each line can be quoted as is, with the source "Vendar Research, November 2026" and a link to this page.

• Of 10,532 documented claims about 395 US SEO and app development agencies, 10% come from anyone other than the agency itself.
• 91% of agency claims can be found on the page they cite. Only 3.8% of measured results come from an independent source.
• 45% of agencies have no independent claim on file at all.
• 92% of agencies name a client. 9% have a measured result published by someone other than themselves.
• Half of the agencies (53%) publish any price. Among mobile app developers, 8% do.
• For Flutter development, International SEO and React Native development, none of the ranking positions across six cities rests on an independent source.
• 18% of agencies do not qualify for any of 156 need-specific rankings because none of their documented work matches a buyer need.

Download the data (CSV): one row per agency with its counts of claims, independent claims, source sites, documented work, named clients, measured results, whether it publishes a price, and how many rankings it appears in. No claim texts. Licensed CC BY 4.0.

Cite as: Vendar Research, "Who Says So? The evidence behind 395 US SEO and app agencies", 2026-11-02, https://www.vendar.org/blog/evidence-gap-2026

Questions, a cut for your city or topic, or a correction: hello@vendar.org.