Which GEO and AEO tactics actually have evidence behind them?
Out of 80 tactics mentioned by enough publishers to compare, 30 are both recommended and backed by measurements, while 14 are openly contested and 12 are mostly called useless or risky. We built this table by saving 435 guides, studies, forum threads and official docs from 320 publishers and checking every quote word for word. Counts do not prove a tactic works, because many writers in this field sell GEO services or tracking tools.
Updated 19 November 2026
Across the 80 tactics with enough sources to compare, 30 are recommended and backed by published measurements. Another 24 are widely recommended but barely measured, 14 draw real pushback or failed tests, and 12 are mostly called useless or risky. So much of the standard AEO and GEO advice is repeated more often than it is tested. Even a tactic labelled measured usually rests on agency case studies or observational data, not controlled experiments.
How we label each tactic
"Recommended and measured": publishers against are at most a quarter of publishers for, and at least three publishers measured it working.
"Contested": publishers against are more than a quarter of publishers for.
"Recommended, little measured": few publishers are against it, but fewer than three measured it working.
"Mostly called useless or risky": more publishers are against the tactic than for it.
Expired domains, parasite pages, paid link schemes
0
1
8
0
0
Mostly called useless or risky
How an agent picks from a list
Step
Recommend it
Measured it working
Call it useless or warn
Measured no effect
Measured harm
Our label
Position in the list the agent reads
0
1
0
2
1
Mostly called useless or risky
Rating and review count shown to the agent
1
3
0
0
0
Recommended and measured
Price against the other options; same price for agent and human
1
2
0
0
0
Recommended, little measured
Badges and labels on the listing (platform pick, sponsored)
1
1
0
0
1
Contested
Title and description wording: plain facts, the searched words, cues that match the request
3
2
0
1
1
Recommended, little measured
A third-party pick (review site, forum) the agent reads before it shops
1
2
0
0
0
Recommended, little measured
Each model weighs the same factors differently
0
3
0
0
0
Recommended and measured
The listing or booking platform blocks agents, so its businesses are unreachable
1
0
2
0
0
Mostly called useless or risky
Persuasion text, fabricated authority and social-proof claims
0
2
0
2
0
Contested
The business's own shopping or booking assistant on its site
0
3
0
0
0
Recommended and measured
How often agents complete a purchase today
4
0
0
0
0
Recommended, little measured
We saved 435 guides, studies, forum threads and platform docs from 320 publishers in full, recorded 1592 word-for-word method statements across 99 tactics, and logged 216 reported results with numbers. Every count is by publisher, not by article, so one prolific blog only gets one vote per stance, and every quote is checked word for word against our saved copy. A high count is not proof that a tactic works: many GEO writers sell GEO services or software, and a measured claim usually means one site's before-and-after numbers rather than a controlled test.
Common questions
Does a high publisher count prove a GEO tactic works?
No, a high count only shows what writers in the field agree on right now. Many people publishing AEO and GEO advice sell GEO services or tracking software, and many repeat the same tips without testing them. Look at the measured columns and read the word-for-word quotes to see whether anyone ran a controlled test or just watched one site's traffic.