AI Summary
When shoppers ask ChatGPT for product recommendations, does your brand show up? For 90% of brands, the answer is no. We ran 10,000 shopping queries across hundreds of product categories to understand why a small group of brands dominates AI recommendations while everyone else stays invisible.
The pattern: about 10% of brands appear in half or more of relevant queries. The rest, including brands with better products, stronger reviews, and bigger budgets, barely register.
What separates winners from the invisible majority? Citation portfolio strategy. After tracking every brand ChatGPT mentioned and every source it cited, we found measurable patterns: category specialists beat major media 2X, concentration beats distribution, Reddit punches 5X above its weight, and your own website could drive 8X more value than it currently does. The brands building strategic citation portfolios today will dominate AI recommendations for years.
Your customer opens ChatGPT. Asks questions about a product or category they are interested in.
Does your brand show up?
Probably not.
We ran 10,000 ChatGPT shopping queries over six months. Basic stuff customers type every day. "Best X." "X for beginners." "Affordable whatever." What we found wasn't encouraging for most brands. Actually, it was devastating for about 90% of them.
Here's the shape of it: maybe one in ten brands shows up in half the recommendations or more. The rest are ghosts. Brands with better products, stronger reviews, bigger marketing budgets, just absent from the conversation entirely.
The winners aren't random. They built something we started calling citation portfolios, though that makes it sound more intentional than it often is. Some stumbled into it. Some built it deliberately. Either way, they show up 50-90% of the time while everyone else fights over scraps.
We tracked every brand ChatGPT mentioned across these 10,000 queries. Every source it cited. Every pattern in how it decided who to recommend. The portfolios told a story about trust that most brands don't understand yet.
The method was straightforward, if tedious.
ChatGPT gives you 3-5 brand names per query. Not a list of 20. Not "here are all your options." Three to five. That's the game.
We tracked hundreds of product categories. "Best product" and the long-tail stuff. "Best product for [customer type] in [location]." The searches people actually do at 11pm when they're trying to solve a problem.
In every single category, 1-2 brands owned the top slots. Not always the brands ranking #1 on Google, which was the first surprise. Different pattern entirely. These brands appeared 50-90% of the time. The math is brutal: if ChatGPT recommends 3-5 brands and 1-2 of them are always the same, everyone else is competing for 2-3 spots across millions of queries.
We analyzed the sources ChatGPT cited. Their coverage patterns. How often they updated. Which brands they featured and in what context. The portfolios that emerged had a structure to them. Not rigid, but visible.
Category Specialists Beat Major Media 2X
When ChatGPT cited category specialists (the publications that live and breathe one product vertical), brands appeared in about 20-25% of those citations.
When it cited major media, brands appeared 10-12% of the time.
One citation from the right specialist outlet had roughly double the impact of a Forbes or Wall Street Journal mention.
The niche site with 50,000 readers beat the publication with 5 million readers. Every time. ChatGPT doesn't care about your media kit's circulation numbers. It cares about domain expertise, about whether you're the publication that actually understands the product category at a technical level.
Here's what that means: specialists first. Major media if your budget allows it or if you need it for other reasons (fundraising, M&A, whatever). But specialists alone can get you to 50% visibility. We watched it happen.
Concentration Beats Distribution
We found brands with what looked like good citation mixes. Some major media. Some specialists. Reasonable balance across source types.
They still underperformed.
The missing variable: depth.
Better to own 2-3 specialist sources with 10 citations each than to have 10 sources with one mention apiece. Same total number of citations. Completely different impact.
The pattern held across categories. ChatGPT trusts brands that key authorities mention repeatedly. In buying guides, reviews, comparison posts, annual roundups, how-to content. Multiple contexts. Multiple citations. Deep coverage.
Getting mentioned everywhere once looks good on a media report. It doesn't move the needle with AI.
Your Website Could Be Driving 8X More Value
Average brand: 5.5% self-citations (their own site getting cited by ChatGPT).
Top performers: 15-45% self-citations.
That's not 2X better. That's 8X.
The gap comes down to content structure more than volume. ChatGPT is looking for useful answers in a non-biased format. Information that helps users, not marketing that helps brands. Create that kind of content and you win mentions.
It's the only citation channel you control completely. Most brands are leaving nearly an order of magnitude of value on the table because they don't understand what makes content cite-worthy to an AI.
Reddit Punches 5X Above Its Weight
Reddit represented 2.7% of all citations we tracked.
It appeared in 12.8% of ChatGPT's responses.
Nearly five times more reach per citation than average.
The efficiency comes from how ChatGPT weights community consensus. A single Reddit thread can validate multiple brands at the same time through authentic user discussion. Real people comparing products, sharing experiences, arguing about trade-offs. It cuts through commercial noise in a way that polished marketing never will.
Brands investing 5-10 hours per month in genuine community participation (actually being helpful, not promoting) see visibility comparable to five-figure PR campaigns. The ROI isn't even close.
Multiple Paths Work (You Have To Pick One)
The brands hitting 50% visibility or better aren't all following the same playbook. We found at least three distinct approaches.
Different mixes of owned content, major publications, specialist coverage, community presence, tier-3 sources. No single formula. Different brands, different paths, same result.
What they share: strategic allocation, concentrated depth, active management. They're not hoping to get cited. They're building citation portfolios the way investors build stock portfolios. With intention, with rebalancing, with an understanding of what drives returns.
Pick your path based on budget, brand recognition, business needs. But pick one. Hoping for the best isn't a strategy.
This Isn't Future-Tense Anymore
AI shopping is already here.
When ChatGPT recommends products, people buy based on those recommendations. The data is early but directional: this influences purchase decisions.
Winners are building moats right now. Their visibility compounds over time. More citations lead to more recommendations lead to more citations. It's a loop.
Everyone else will fight for scraps through higher CPCs and endless SEO optimization cycles, watching traffic slowly migrate to AI interfaces they don't show up in.
The time to invest in AEO is now, while the patterns are still visible and the playbooks still work. Once you're in ChatGPT's trusted set, you tend to stay there. Until someone earns that trust away from you.
What It Actually Means
Ten thousand queries later, the insight is almost boring in its simplicity.
Visibility in ChatGPT isn't about who spends the most. It's about who earns consistent trust from the right sources.
The patterns are measurable. The paths are proven. The brands building strategic citation portfolios today will probably dominate AI recommendations for years.
The rest will wonder why they're invisible while their competitors are everywhere.
FAQs
What is a citation portfolio?
A citation portfolio is the mix of sources (specialist publications, major media, community forums, your own site, etc.) that mention your brand and products. ChatGPT uses these citations to decide which brands to recommend. Think of it like an investment portfolio: the right allocation and concentrated depth matter more than total volume.
Why does ChatGPT recommend some brands over others?
ChatGPT looks for consensus across trusted sources. Brands that get mentioned repeatedly by category specialists, have distributed reviews across multiple platforms, and create cite-worthy content on their own sites trigger consistent recommendations. It's not about who spends the most on marketing. It's about who earns trust from sources ChatGPT weights heavily.
Do I need to be on Wikipedia to show up in ChatGPT?
No. Wikipedia is the most-cited source overall (24% of queries), but individual brands only get mentioned in 2.5% of those citations. What Wikipedia does is validate category legitimacy. Brands with Wikipedia pages appear 3X more often overall because the page acts as infrastructure, not a direct driver. You can hit 50%+ visibility without Wikipedia if you have the right citation mix.
What's more valuable: major media or specialist publications?
Category specialists deliver 2X the impact of major media. When ChatGPT cites a specialist publication, brands appear 20-25% of the time. Major media citations result in 10-12% appearance rates. One citation from the right 50K-reader specialist site beats a Forbes mention. ChatGPT values domain expertise over audience reach.
How many sources do I need?
Six is the threshold. Brands with 5 diverse, high-quality sources see 25-35% visibility. Hit 6 sources and visibility jumps to 60-70%. Beyond 10 sources, returns diminish rapidly. But quality matters: 6 wrong sources still leaves you at 20-30% visibility. It's not just about hitting the number, it's about the right mix.
Should I spread citations across many sources or concentrate on a few?
Concentrate. Brands dominating 2-3 specialist sources with 10 citations each see 50-80% visibility. Brands scattered across 10 sources with one mention each see 15-30% visibility. Same total citations, completely different outcomes. ChatGPT trusts brands that key authorities mention repeatedly across multiple contexts (reviews, guides, comparisons, roundups).
Can I cite my own website?
Yes, and you should. Top performers get 15-45% of their citations from their own sites. Average brands get 5.5%. That's an 8X gap. The key is creating useful, non-biased content that answers real questions. ChatGPT looks for information that helps users, not marketing that helps brands. It's the only citation channel you control completely.
Does Reddit really matter for AI recommendations?
Yes. Reddit represents 2.7% of all citations but appears in 12.8% of ChatGPT responses. That's nearly 5X more reach per citation than average. Why? Community consensus is weighted heavily. A single thread can validate multiple brands through authentic discussion. Brands investing 5-10 hours monthly in genuine participation see visibility comparable to five-figure PR spends.
Do reviews on my site count, or do they need to be on third-party platforms?
Third-party platforms matter more. ChatGPT aggregates reviews across multiple sources and explicitly states its ratings don't match any single marketplace. Distribution beats dominance: 100 reviews across 5 platforms creates stronger consensus than 500 reviews on one platform. Optimal distribution: 35-40% on your primary platform, then 25-30%, 20-25%, and 10-15% across others.
How long does it take to see results?
6-12 months for meaningful visibility. Specialist outreach takes 3-6 months to secure initial coverage. Building concentration (5-10 citations per source) takes another 6-9 months. Community presence requires 6-12 months of authentic participation. Quick wins exist (technical fixes, owned content optimization), but durable visibility requires sustained effort. The advantage: once you're in ChatGPT's trusted set, you tend to stay there.
What if I have a limited budget?
Focus on Path A: Specialist-First. You can hit 50%+ visibility with zero major media spend. One brand in our study achieved 50% visibility with 0% major media citations. Concentrate on 2-3 category specialists, invest 5-10 hours monthly in community participation, and optimize your owned content. This approach works for unknown brands with limited resources.
Is this just SEO with a different name?
No. The strategies overlap but the signals are different. Google rewards backlink volume and domain authority. ChatGPT rewards citation depth and source quality. Google wants lots of sites linking to you. ChatGPT wants trusted authorities citing you repeatedly. You can rank #1 on Google and be invisible in ChatGPT if you don't have the right citation portfolio.
What's OAI-SearchBot and why does it matter?
OAI-SearchBot is OpenAI's web crawler. If your robots.txt blocks it, ChatGPT can't see your site at all. This is an instant invisibility trigger. Check your robots.txt file and make sure OAI-SearchBot is allowed. It's used exclusively for search results, not model training. Without it, you're invisible no matter how good your content is.
Can I pay to get recommended by ChatGPT?
No. There's no paid placement in ChatGPT recommendations currently. Visibility comes from earning citations from sources ChatGPT trusts. You can pay for PR to get media coverage, you can invest in content creation, but you can't buy your way into recommendations directly. This is actually good news: it means smaller brands can compete if they build the right citation portfolio strategically.
What if my competitors already dominate my category?
Categories aren't locked yet, but the window is narrowing. ChatGPT typically recommends 3-5 brands per query. If 1-2 brands already claim top slots consistently, you're competing for 2-3 remaining positions across millions of queries. The math gets harder but not impossible. Focus on concentration with specialists, build owned content assets, and invest in community presence. First movers have an advantage, but patterns are replicable.
Do I need different strategies for Google vs. ChatGPT?
Yes and no. Technical foundations overlap: structured data, clean URLs, quality content. But the ranking signals differ. Google values backlink quantity and domain authority. ChatGPT values citation depth and consensus. Google wants 100 sites linking to you once. ChatGPT wants 6 trusted sources citing you 5-10 times each. You need both strategies, but they're not the same strategy.
How do I know if my citation portfolio is working?
Track visibility rate: what percentage of relevant queries result in your brand being recommended. Test 20-50 queries monthly that match your products. Track which sources ChatGPT cites when it recommends you (or doesn't). Monitor competitor movement. Measure citation count, source diversity, and portfolio allocation. If you're below 15% visibility, your portfolio isn't working. 50%+ visibility means you've built something durable.
Should I hire someone or do this in-house?
Depends on resources and expertise. Technical fixes (robots.txt, schema, content structure) can be done in-house with guidance. Specialist outreach requires relationships and credibility in your category. Community participation needs authentic engagement, hard to outsource. Portfolio strategy and allocation decisions benefit from experience. Many brands start with audit and strategy from specialists, then execute some pieces in-house while outsourcing others.
What happens if I don't invest in AEO?
You stay invisible as shopping behavior shifts to AI interfaces. Winners are building moats now. Their visibility compounds over time: more citations lead to more recommendations lead to more citations. Everyone else fights for scraps through higher CPCs and endless SEO cycles, watching traffic slowly migrate to AI platforms they don't show up in. The time to invest is while patterns are still visible and categories aren't locked.
Ready to transform your digital strategy
Learn how AI search can revolutionize your brand's online visibility and engagement.
More insights on AI search
Discover cutting-edge strategies for digital marketing success
Stay ahead in AI search
Subscribe to our blog and get insights before anyone else