The shortlist effect: why only a few brands
When you ask ChatGPT for a recommendation, it typically names three to five brands—not because it knows of only those, but because that's what emerges from how it processes corroboration and relevance in its training data. The shortlist effect is the natural result of how many brands meet the threshold of clarity, corroboration, and fit to the specific question you've asked.
A brand enters the shortlist when it crosses a corroboration threshold in the training data. Dozens of independent citations across trusted sources carry more weight than a single strong mention. This creates a hierarchy: the most frequently cited and consistently described brands rise to the top; others remain known to ChatGPT but don't make the cut. The shortlist isn't fixed—different questions yield different lists—but the pattern is consistent.
Category leaders, corroboration, and crowding
Category leaders occupy the shortlist almost by default. In any market, a few brands are discussed far more often than others, and when multiple reputable sources cite them consistently, they become the anchors of the category. A brand like Figma in design tools or Notion in productivity has accumulated enough corroboration across authoritative sources that ChatGPT names them reliably.
Corroboration creates a crowding effect. As more brands meet the basic threshold, the most-corroborated ones push others down. A lesser-known brand might have one strong mention in a niche publication, but if five competitors have multiple mentions across mainstream publications and specialist sites, those five will dominate the shortlist. This is not deliberate exclusion; it's an emergent result of the pattern.
| Query type | Brands typically named | Who makes the cut |
|---|---|---|
| Broad 'best X' list | 4–6 | Category leaders and the most-corroborated alternatives |
| Single-best recommendation | 1–3 | The market leader, plus maybe one clear alternative |
| Niche or specialized query | 2–4 | Specialists recognized in that niche, plus general-purpose options |
| Comparison query | 3–5 | Direct competitors mentioned together in the sources |
| Constraint-based query (budget, use case) | 2–4 | Options that fit the constraint; leaders in that tier |
How query type affects shortlist size and composition
How query type shapes the shortlist
The number of brands that make the shortlist depends on what you're asking. A vague question like 'What are the best design tools?' invites a broader shortlist because you're implicitly asking for multiple options. A more focused question like 'What's the best design tool for beginners?' narrows the focus and often yields fewer recommendations, even though ChatGPT knows of many tools. The specificity of your question governs the selectivity of the response.
Category-associated clarity also determines shortlist entry. A brand that is unambiguously 'the design tool for UI work' or 'the CRM for nonprofits' has an inherent advantage over a brand that could serve multiple purposes. When ChatGPT can confidently associate a brand with a specific category or use case, it becomes shortlist eligible. Ambiguous or multi-category brands must overcome that with even stronger corroboration to compete for a slot.
- Broad 'best X' queries typically name 4–6 options; narrower 'best X for Y' queries name 2–4
- Single-best queries heavily favor the market leader; alternatives rarely make the cut
- Niche queries can favor lesser-known specialists if they're strongly associated with that niche
- Comparison queries name all brands explicitly mentioned in the question, plus other corroborated alternatives