86%
of plain runs for personal injury lawyer came back altered. All 50 metros were altered at least once in two runs.
50 US metros, 100 plain runs and 50 repaired runs, DataForSEO Google results API, November 21, 2026
49 of 60
plain runs were altered across six trades in ten cities. Dental implants was altered in 10 of 10.
30 searches each run twice plain and once repaired, November 21, 2026
0
was the usual number of sites an altered first 10 shared with the repaired first 10.
Median over 100 plain runs for personal injury lawyer, November 21, 2026
What we ran and how we counted
We used DataForSEO's Google results API, the same kind of source many rank tracking tools are built on. Each search was sent with the location set to the middle of the city, on desktop, in English, and we kept the first 10 ordinary results.
Every search ran three times within a few minutes. Two runs were plain. The third used the provider's repair, released November 16, which checks each result against Google's Web view and replaces the ones that do not match.
A plain run counts as altered when fewer than 3 of its first 10 sites appear in the repaired first 10. Google reorders results all the time, so we set the bar low on purpose. A run that only shuffled the same sites does not count.
Every metro, most of the time
For personal injury lawyer, 86 of 100 plain runs were altered. In 36 metros both plain runs were altered, in 14 one was, and in none were both clean. It comes and goes between runs a few minutes apart, so a single clean check tells you little.
The altered pages are not random. The sites that showed up most in their first three places were Expertise.com, personalinjurylaw.org, lawyerup.co, lawyersforhire.com and bpublic.com. In the repaired runs the first three places went to Super Lawyers, Justia and local firms.
- 01
Dallas
Plain run, first places, yellowpages.net, mighty.com and dallas.opendi.us. Repaired run, mullenandmullen.com, Super Lawyers and bencrump.com.
- 02
New Orleans
Plain run, guysworld.in, lawyersforhire.com and americantopattorneys.com. Repaired run, Justia, morrisbart.com and bencrump.com.
- 03
Pittsburgh
Plain run, lawyerup.co, Martindale and Yelp. Repaired run, edgarsnyder.com, Super Lawyers and ssf-lawfirm.com.
It is not only law
We ran six kinds of local search in ten cities, from Atlanta to Boise, the same three times each. 49 of 60 plain runs were altered.
- 01
Dental implants
10 of 10 plain runs altered.
- 02
Personal injury lawyer
9 of 10.
- 03
HVAC repair
9 of 10.
- 04
Emergency plumber
8 of 10.
- 05
Med spa
7 of 10.
- 06
Roof replacement
6 of 10.
Agencies' own searches too
We also ran law firm marketing agency in the same 50 metros. Of 82 plain runs that returned a usable page, 72 were altered. The altered pages led with agency directories and list sites such as Semrush's agency directory, webtonic.io and topseobrands.com. The repaired pages led with agencies, among them milemarkmedia.com, ilawyermarketing.com and modfxmedia.com.
So an agency checking its own place for the search its buyers type can be looking at a page no buyer sees.
How we know which page is the real one
We did not sit a person in each city. Three things point the same way. The repaired pages are checked against Google's own Web view. They are full of the kind of result people expect, law firms for a lawyer search, where the altered pages are full of directories most people have never seen on a first page.
And for one business we work with, Google's own Search Console reports an average place of about 4 for a search where the plain run did not list the business in the first 30 results and the repaired run listed it first.
What to do if you report rankings
- 01
Compare your tracker with Search Console
Search Console is Google's own record of where your pages showed for real searchers. If your tracker says you fell off the first page and Search Console shows steady impressions and place, believe Search Console.
- 02
Ask your rank tracking tool what it changed
Ask whether it detects altered results and what it does when it finds them. A tool that changed nothing in November may be reporting pages nobody sees.
- 03
Do not act on one check
The same search came back clean and altered minutes apart. Before you tell a client they dropped, run it again and check the sites in the first places. A first page of directories is the sign.
- 04
Look again at reports from the last few weeks
We do not know when this began for each search. A sudden fall across many keywords this autumn may be the measurement and not the site.
What we do not know yet
This is one provider on one day. We have not tested other providers or the tracking tools built on them, and some may already correct for it. We treat the repaired page as the real one for the reasons above, and we checked it against Search Console for one business only.
For law firm marketing agency, 18 of 100 plain runs returned too few results to count. We will run the same searches again in December and report whether the share changes.
Questions we get asked
Are Google rank tracker results wrong right now?
- Some are. On November 21, 2026 we ran personal injury lawyer in 50 US metros through DataForSEO's Google results API. 86 of 100 plain runs returned a first page that shared fewer than 3 sites with the repaired page. We tested one provider, so we cannot say which tracking tools are affected.
What are poisoned Google results?
- It is the name search data providers use for altered pages Google returns to automated requests. In our runs the altered pages listed directories and list sites where the repaired pages listed local businesses.
How can I tell if my rank report is affected?
- Compare it with Google Search Console, which records where your pages showed for real searchers. Then look at the sites your tracker lists in the first places. If they are directories you have never seen on a first page, the check was probably altered.
Does this affect what real people see on Google?
- We found no sign of that. The altered pages went to automated requests. The repaired pages match Google's Web view, and for one business we checked they match Search Console.
Does this affect AI assistants?
- We did not test that here. This finding covers Google's ordinary results as read by a data provider.
Your free report
See what Google and the AI assistants show for your business.
Get a free pick report. We ask Google and AI assistants your customers' questions in your town and show who they named and what to fix first.
- 01We read your site and work out what your customers ask an assistant for.
- 02We run the top assistants on those requests in your area.
- 03You get where you rank, whether they would book you and why, and what to change first.