In short: Five objections come up almost every time. Three have straight answers. Two are correct, and conceding them is what makes the other three believable.
The two that are right: you cannot attribute revenue to this yet, and a lot of what is sold as GEO is repackaged SEO. Argue with either and you are defending the part of the field that deserves the scepticism.
An objection is a question about risk. Answer the risk rather than the sentence.
"There Is No Traffic From It"
Usually true, and usually measured with an instrument that cannot see the thing.
A mention with no link sends nobody, and it still put your name in front of somebody who asked. Analytics built on sessions records that as zero, so the absence of traffic is not evidence of absence of value: it is the expected reading either way, which makes it useless as a test.
The answer is not to claim hidden value. It is to say what can be counted instead, which is presence across a fixed question set, and to agree in advance that this is the number being judged rather than sessions.
"We Cannot Measure It, So We Cannot Manage It"
Half right, and the half that is wrong is the important one.
You cannot measure it precisely. Answers vary between identical runs, so a single reading is an anecdote. But a frequency across repeated runs, on a fixed question set, compared against how much the same question disagrees with its own rerun, is a measurement, and it is more honest than a rank was.
Measuring Your Own Noise Floor is the whole answer to this objection, and the person raising it is usually the person most convinced by it, because it concedes the imprecision first and then handles it.
"This Is Just SEO With A New Name"
Concede most of this. It is the objection with the most truth in it and the one most often argued with badly.
Crawlability, structure, page speed, authorship and earned coverage all carry over unchanged. Somebody who has done technical SEO well for a decade has done most of this already, and telling them otherwise is both false and insulting.
What is actually different is a short list: inclusion instead of position, value arriving without a click, selection discarding material before merit is judged, the system searching its own rewritten query, results varying between identical runs, and platforms that disagree about which sources exist. What's Different is that list.
The honest framing: this is not a new discipline, it is a change in what the existing discipline is optimising for, plus one genuinely new problem, which is being described as part of your category by other people in places these systems read. Said that way it stops sounding like a land grab and starts sounding like work.
"It Is Too Early"
The answer is not that it is late. It is that the slow part does not wait.
Reachability and extraction are days and weeks of work. Association is quarters, because it is built by being written about by other people, and none of that starts until somebody starts it. A company that begins when the volume arrives begins the slow part at the point it needed to have finished.
That is an argument for starting small rather than for starting big, which is the next lesson and the reason this objection is easier to answer than it looks.
"Prove The Return First"
Concede this too, precisely. There is no defensible session-based return for a mention nobody clicked, and any number offered is invented. What GEO Is Worth, Honestly is why.
What can be offered instead is a cost and a decision rule: this is what it costs, this is what we will measure, and this is the result that would make us stop. A person asking for proof of return is usually asking how to limit the downside, and a stated stopping condition answers that better than a forecast would.
The objection nobody says out loud: whoever owns search is being asked to fund something that might make their own numbers look worse, because traffic can fall while presence rises. Naming that in the room, without blame, converts the most likely quiet opponent into the person whose reporting problem you are solving.
What Not To Do With An Objection
Three failures, in order of how much damage they do.
Do not answer with a forecast. It moves the argument to ground where you have no evidence and the other person has as much standing as you do.
Do not promise a number you cannot control. A target for a rank or a blended score buys agreement now and loses the next three meetings, which is the subject of Targets You Cannot Set.
Do not treat a correct objection as resistance. Somebody who says the return is unprovable is right, and handling it as an obstacle tells them you have not understood your own field.
Key Takeaways
- Five objections recur; two of them are correct, and conceding those is what makes your answers to the rest credible.
- No traffic is the expected reading whether or not there is value, so it cannot function as a test. Agree on presence across a fixed question set instead.
- Concede that most of GEO is SEO. What is genuinely different is a short, nameable list.
- Too early is answered by the slow step: association moves in quarters and does not start until somebody starts it.
- There is no defensible return figure. Offer a cost and a stated condition that would make you stop, which is what the question is usually really asking for.
Check yourself
Before you move on
Not scored, not recorded, and not part of the certificate. Both answers are settled by a sentence in this lesson, and the reasoning appears whichever option you pick.
- 01
Somebody says there is no traffic from AI answers, so the work is not worth doing. What is wrong with that as a test?
- 02
Which two objections should you concede rather than argue with?