AEO/GEO: Your AI citations are climbing, but not your revenue? Here’s why

2,800 AI citations. Revenue didn’t budge.

March 2025. An e-commerce site with 3,200 SKUs. 45,000€ in monthly revenue. The marketing manager was thrilled: citations in Google AI Overviews and ChatGPT climbed from 1,200 to 2,800 in three months.

But the sales dashboard stayed flat.

Every month, 45,000€. Not a euro more. While mentions skyrocketed, the conversion rate plateaued at 0.4%. The bots loved the brand. Humans, not so much.

I see this with one client in three. Teams celebrate citations like trophies. Yet revenue stays locked. That’s the GEO trap: the dopamine rush of mentions without the financial reward.

I dug into the problem with this client.

We listed the citations. 2,800 occurrences were about informational questions, generic definitions, knowledge-seeking queries: « what is X », « how does Y work ». None of them pushed toward purchase. None triggered a click with transactional intent.

Result? Ghost traffic. Useless visibility. The frustrating situation: you’re seen, but not bought.

Chasing citation volume is a false lead. What matters is what the citation triggers.

Why your brain loves these citations (and your accountant hates them)

When your citation graph climbs, your brain releases dopamine. It’s the same molecule as a social media like. I watch clients hooked by that feeling of forward momentum. Their SEO effort is paying off, yes. But dopamine has nothing to do with profit.

Guillaume Attias says it in the DOSE training at BMO Academy: dopamine rewards immediate action, not long-term value creation. You feel good seeing your name on ChatGPT, but that hit doesn’t fill the till.

This neurochemical trap is called salience bias. You overweight what you measure easily (mentions) and ignore what you don’t (purchase intent). Result: your margins stagnate.

In AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization), the trap amplifies. Tracking tools bombard you with pretty dashboards: « +87% visibility on ChatGPT », « +120 citations on Google AI Overviews this month ». None tell you how many shopping carts those mentions triggered.

Your accountant is right. A non-transactional citation costs you. Time spent writing content that doesn’t sell. Resources poured into feeding bots instead of customers.

The real question isn’t « how many citations? » but « how many sales per citation? ». Once you ask that, the whole picture shifts.

What NP Digital’s study of 100 AEO/GEO campaigns reveals

I’m not reinventing the wheel. Neil Patel and his team at NP Digital just combed through over 100 AEO and GEO campaigns. Goal: understand why brands pile up citations without seeing revenue move. Their conclusion is sharp: most marketing teams confuse AI visibility with commercial performance.

In his blog post, Patel lists common errors. I reframe them from my field perspective:

• Error #1: no post-click tracking. Clicks from AI Overviews or ChatGPT drown in direct traffic. Without specific UTMs, you can’t tell if a sale came from an AI citation. You’re flying blind.
• Error #2: optimizing for bots, not buyers. Many pieces are built to please answer engine algorithms: packed with definitions, thin on product reassurance. Result: the bot cites you, but humans don’t convert.
• Error #3: confusing brand awareness with conversion. A brand citation in an AI Overview can boost image, sure. But if it’s not tied to a transactional query, it drives zero sales. That’s a branding investment, not performance.

NP Digital’s analysis highlights a number every e-commerce marketer should sit with: 73% of AEO/GEO campaigns aren’t linked to a measurable revenue goal. You’re optimizing for AI, not business. It’s like opening a store and waiting for people to walk in without ever showing them products.

Patel insists: brands winning in GEO shifted focus from mentions to high-intent queries. They don’t chase citations everywhere, only on questions that precede a purchase. A total perspective shift.

The simple method to link AI visibility and sales

Back to our 45,000€-a-month e-commerce site. We applied a three-step method, inspired by Patel’s lessons and my cocon practice.

Step 1: tag each citation with a UTM that tells its intent. On all pages that AI can cite (and those already cited), we added precise UTM parameters: source (chatgpt, google-ai-overview), medium (organic-ia), content (informational vs transactional). One subtlety: we split informative landing pages from product pages. For every citation, we know where the visitor came from and which content type triggered it.

Step 2: prioritize high-intent queries. We know semantic cocoons, so we spotted buyer questions just before purchase: « X vs Y comparison », « reviews of X », « best price for Y », « where to buy Z ». These are the targets for content we push to AI. Out with « everything about X » articles. In with comparison pages, reassurance pages, product FAQs. Each piece answers a buying question AND contains a natural link to a product page with UTM.

Step 3: measure closed-loop attribution. We connected AI UTM sessions to the conversion funnel with Google Analytics 4. Within a month, we saw the signal: transactional clicks converted 5–7x more than informational clicks. Bingo. We killed content that brought clicks without conversion. We redeployed resources to profitable content.

This method costs zero extra tools. Just rigor. And it changes the game. We don’t talk about « AI mentions » anymore—we talk about « AI acquisition channels ». Suddenly, GEO becomes a revenue engine.

This chart compares key metrics before and after our targeted optimization. Notice how fewer citations can drive more revenue when they are transactionally aligned.

Fewer citations, more revenue: the profitable paradox

Six months into this method, results landed.

Citations in AI Overviews and ChatGPT dropped from 2,800 to 1,500 per month. Yes, a 47% decline. But at the same time, the conversion rate per citation jumped from 0.4% to 1.2%. Monthly revenue grew from 45,000€ to 96,000€. That’s a 220% jump.

This paradox shows relevance beats volume. You can lose half your citations if the remaining ones drive three times more sales. Growth isn’t about quantity—it’s about calibration.

I often see the same reaction from e-commerce teams: fear of watching mentions drop. Yet once they accept it, the financial gain flips their thinking fast. SEO ego fades against the bottom line.

One number struck me in this case: of the 1,500 remaining citations, 900 were tied to identified transactional queries. The ratio flipped from where it started. That rebalance explains the performance.

The classic mistake is piling onto what already exists. Real optimization is subtraction. Prune content that doesn’t fuel sales. Concentrate effort on the 20% of citations generating 80% of revenue. Nothing new under the Pareto sun, but the message struggles to land in AEO/GEO.

Will your next citation be worth a customer, or just a line in a dashboard?

The answer is in your hands. When you celebrate your latest ChatGPT citation, ask first: « What’s the odds this visitor buys? » If it’s nil, what are you really celebrating?

AI visibility is a means. Its value? Serving a conversion machine. A system that runs from first click to checkout. With my clients, I forge architecture where every citation is a call to act, a line that counts on the ledger.

So next time you open your AI citation dashboard, ask: « What share of these mentions triggered a sale this month? » If you don’t know, you’ve got a blind spot. But more than that—a real opportunity.

Your GEO strategy exists to sell. The day you align your content to that aim, you’ll see citations in a new light.

Stéphane Jambu

SEO & AI Engineer

I build growth systems / AI / Neuroscience | 650+ clients · 80 LinkedIn testimonials · 30 years of expertise · 15 years of systems running without me.

Follow on LinkedIn