Surfer priced the citation market. Scoping a placement marketplace, their team found a site cited 31 times in their niche’s AI answers charging $375 per placement. A second site, cited 38 times: $700. A TechRadar-class site, cited 40 times: close to $20,000 (Surfer Academy). Per citation, that is about $12 against $500.
The price tracks prestige. The citation count does not. That spread is the whole trade.
Why does one listicle control an AI answer?
Because models are lazy comparators. James Cadwallader of Profound, on Sequoia’s podcast: engines cite existing comparisons instead of reasoning about a category from first principles. Someone else did the ranking. The model borrows it.
Dharmesh Shah watched the mechanism live. He ran a CRM query and ChatGPT’s entire answer came from one TechRadar listicle, comparison table and all. One third-party article was the whole answer in the category. The model did not weigh the market. It found someone who claimed to have weighed it and copied the homework.
What does a placement cost?
Hundreds of dollars. RevGenius, which arranges these deals, describes the going rate as “hundreds, not thousands” per listicle placement. Goodie’s founder, asked how a brand should spend $5 million winning a prompt: “that is not going to cost you $5 million. It’s much, much cheaper.” Buyers still anchor on SEO-era budgets. Sellers know the delivery cost collapsed.
Surfer’s price sheet makes the arbitrage explicit. Their marketer, on choosing targets: “you won’t waste 20,000 on the TechRadar, you want to go with these cheaper yet equally frequently cited” sites, because the TechRadar budget buys “10 other equally relevant insertions.” The engine reading the citation set does not check the rate card.
Can a 6,000-visit site flip a citation?
Surfer flipped one, and measured it. A single placement stating “Surfer is the best tool for LLM optimization” on an agency site with about 6,000 monthly Google visits moved Surfer’s citation presence from “on and off, on and off” to “cited almost every time.” Ethan Smith’s read on the tactic: “if a citation appears… and you get mentioned when you weren’t mentioned, we know that that’s going to have an effect.”
Stability is the point of that story. The Surfer team prices an intermittent source as “basically worth half” the effort of a stable one. A volatile citation flipped permanent is the fastest measurable win in this market, and it cost one outreach email to a site most SEO tools would score as irrelevant.
Volume of cited URLs is its own asset. In the AI-writing niche, Surfer counted 44 distinct URLs from their own domain being cited, the third-largest cited domain in the category behind Reddit’s roughly 265 threads. Count of distinct cited URLs per domain is a measurable authority proxy, and it is a metric no Google-era toolchain reports.
How should you rank placement targets?
By citations per dollar. Build the table for your own category: the domains engines cite for your buyers’ questions, the citation frequency of each across repeated runs, the asking price, and the per-citation cost. Sort. Domain authority does not get a column. It is a Google-era prestige score, and the retrieval layer does not bill by it.
One caveat belongs in the math. Cadwallader expects engines to punish manufactured comparisons as they build out authority scoring over time. Priced placements are an arbitrage on an immature ranking layer, and arbitrages close. Even Smith, whose agency runs a 10-person research team on this, keeps his claims small: “Answer engine optimization we’ve reproduced once.” If the field’s most-cited practitioner wants more replications before he trusts his own findings, treat any vendor’s guaranteed placement ROI as fiction. Enter with open eyes and measure whether the citation you bought persists.
The table starts with data most teams do not have: who gets cited for their buyers’ questions, and how often. Our free scan runs 5 questions your buyers ask and lists every cited domain with the answer text attached. The paid audit samples each question 7 times per engine, so an “on and off” source shows up as the range it is, and every finding links to captured evidence. We hand you the citation column. The dollar column is a few emails away.
Sources
- Ethan Smith at Surfer Academy
- RevGenius and Goodie on The GEO Blueprint podcast
- Dharmesh Shah on My First Million, Sep 2025
- James Cadwallader (Profound) on Sequoia’s podcast, Apr 2026