Buyer questions for Nike, drafted by four Claude models

CiteLadder's Prompt Builder drafts the buyer questions a brand tracks in AI answers. For this study, the same step ran on Claude Haiku 5.5, Claude Sonnet 5.5, Claude Opus 5.5 and Claude Fable 5.1. Each model got the same instructions, the same Nike business profile and the same 12 planned slots. Each ran at low, medium and high effort, twice per setting. Only the model and the effort changed. The tables below show every question each model wrote. This page describes what can be observed and measured. Judging quality is left to the reader.

What the business profile says Nike sells

The business profile files Nike under athletic footwear and apparel. The business model is a direct-to-consumer product business, and the buyers are consumers. The category terms are athletic shoes, sportswear, running shoes, basketball shoes, sneakers, activewear, training apparel and soccer cleats.

The profile lists four offerings. Each one became a topic:

• Running shoes: road and trail footwear, from daily trainers to max-cushion and carbon-plated race shoes.
• Basketball and court shoes: basketball and tennis footwear, including signature-athlete models and retro high-tops.
• Lifestyle sneakers: casual sneakers based on classic sport silhouettes, with many colorways and custom designs.
• Training and performance apparel: sports bras, leggings, shorts, tops, fleece hoodies and weather-ready jackets.

Soccer cleats appear in the category terms but not as an offering, so no topic covers them. Each offering carries its own attributes, situations and audiences. The slot plan draws on those.

How the 12 slots were planned

Each topic got three slots, and every topic follows the same pattern:

• Slot 1 targets the awareness stage and is built around a situation. Allowed intents: learn, solve, recommend.
• Slot 2 targets the consideration stage and is built around an audience. Allowed intents: solve, recommend, compare, validate.
• Slot 3 targets the decision stage and is built around an attribute. Allowed intents: compare, validate, buy, recommend.

The facet for each slot, in stage order:

• Running shoes: marathon race preparation, road runners, maximum cushioning.
• Basketball and court shoes: limited-edition release drops, tennis players, court traction pattern.
• Lifestyle sneakers: holiday gift giving, parents buying for kids, custom design options.
• Training and performance apparel: gym strength sessions, women athletes, sports bra support level.

All slots sit in the core cohort. Each model produced 6 outputs from this plan, for 72 questions per model.

What is visibly different between the models

Length. The median question is 9 words for Claude Haiku 5.5 and Claude Sonnet 5.5, and 10 words for Claude Opus 5.5 and Claude Fable 5.1. No model wrote a question over 16 words.

Openings. Claude Haiku 5.5 often opens with the intent itself: "Compare", "Recommend" or "Best". It also tends to reuse profile wording, as in "Compare basketball shoes with herringbone traction patterns for indoor courts". Claude Sonnet 5.5 settles on a few openers and keeps them, such as "Best running shoes for marathon training and race day".

First-person framing. All four models write some questions as a buyer describing themselves. Claude Sonnet 5.5 does this in a fixed form for the parents slot: "I'm a parent, what casual sneakers hold up for active kids?" Claude Opus 5.5 and Claude Fable 5.1 often add a habit or a setting, such as playing tennis twice a week or running mostly on pavement.

Framing as a problem. Claude Opus 5.5 sometimes starts from something that went wrong, for example "Alternative to leggings that go see-through during squats, for women athletes". It also adds physical detail, as in "Basketball shoes with the grippiest traction for dusty indoor courts". Claude Fable 5.1 shares much of this wording. Both models wrote "Who lets you design your own custom sneakers online?"

Question marks. Claude Fable 5.1 ends 62.5% of its questions with a question mark, the largest share here. Some of these are search-style phrases with a question mark added: "Best running shoes to train in while preparing for a marathon?" It also wrote one comparison with "vs": "Daily trainer vs tempo shoe for a road runner, which should I buy?"

Specificity. No question from any model names Nike, another studied brand or a place. The specific detail comes from use cases, body types, surfaces and ages.

Effort levels and repeat runs

Claude Sonnet 5.5 changes little across settings. Its low effort run 1 and run 2 are close to identical. Its marathon question repeats at low, medium and high effort. Where it does change, the change is often a single word, such as grip or traction, roads or pavement.

Claude Opus 5.5 is also steady on some slots. Its custom sneaker question appears at every effort level. Other slots shift more, for example the tennis and women athletes slots.

Claude Haiku 5.5 rewords most slots between run 1 and run 2, even at the same setting. A few phrases still recur, such as herringbone traction for the court traction slot.

Claude Fable 5.1 keeps much of its wording across low and medium effort. Its high effort run 2 differs from its other outputs in several slots, including the marathon and road runner slots.

Across these outputs, moving from low to high effort shows no single direction in length or framing. For several slots, run 1 and run 2 at one setting differ about as much as different effort levels do. The output token figures below are per model, not per effort level.

What the code measured

The code checked all 72 questions from each model:

• Median words: 9 for Claude Haiku 5.5 and Claude Sonnet 5.5, 10 for Claude Opus 5.5 and Claude Fable 5.1.
• Over 16 words: 0% for every model.
• Names a studied brand: 0% for every model.
• Joins two asks with "or": 0% for every model.
• Names a place: 0% for every model.
• Ends with a question mark: 37.5% for Claude Haiku 5.5, 50% for Claude Sonnet 5.5, 47.2% for Claude Opus 5.5, 62.5% for Claude Fable 5.1.
• Near-duplicate pairs: 0% for every model.
• Distinct openers: 58.3% for Claude Haiku 5.5, 31.9% for Claude Sonnet 5.5, 37.5% for Claude Opus 5.5, 40.3% for Claude Fable 5.1.
• Mean output tokens: 2417 for Claude Haiku 5.5, 780 for Claude Sonnet 5.5, 798 for Claude Opus 5.5, 802 for Claude Fable 5.1.

Identical questions do repeat across separate runs in the tables. The near-duplicate figure does not reflect those repeats, which suggests it looks at pairs within a single output.

Every question, side by side

Run 1 of each setting. Each row is one planned slot; the model and the effort are the only things that change across a row and down a stage.

Running shoes

Basketball and court shoes

Lifestyle sneakers

Training and performance apparel

What the code measured for Nike

All six outputs per model (three efforts, two runs), 72 questions each.

Frequently asked questions

Track your own buyer questions in AI

CiteLadder tracks how ChatGPT, Gemini, Claude and Google AI Overviews answer the buyer questions for your brand, recording mentions, position, sentiment and competitors. Every result is tied to the recorded answer, so you can read exactly what each model said.

Explore AI Visibility