0 of 18 done
Your ticks are saved in this browser only. Nothing is sent to us.
Prompt set and baseline
- Prompt set agreed with the client: the questions buyers actually ask (30–100+)
- Prompts split by intent: definition, validation, exploration, comparison, ranking, recommendation
- Baseline run on public UIs (ChatGPT, Perplexity, Gemini), min. 4 measurement runs/month
- Baseline records: which brands each answer names, which pages each answer cites
- Presence reported as % and rotation over time — never as a single-shot position
Target selection (work backwards from the answers)
- Cited pages listed per prompt cluster: comparison lists, review roundups, alternatives pages
- Gap matrix built: competitor × intent × category × funnel stage — where competitors are named and we’re not
- Placement targets prioritised where engines already draw from
- Platform mapping: ChatGPT → Reddit/YouTube/forums; Perplexity → comparison/e-commerce; Gemini → gov/edu/authority
Placement rules
- Mention placed only where a reader would expect to find it
- Says only what’s true of the product; positioning consistent everywhere (engines can’t resolve five conflicting descriptions)
- No hidden text, no bulk “best of” shells, no gaming the answer
- Mention with or without link decided per placement (a mention without a link still feeds engines)
- Same site qualification rules as every placement (DR 40+, traffic, relevance, no link farms)
Re-check and reporting
- Same prompt set re-run each cycle; report named/not-named per prompt
- Track across 3+ platforms in parallel (brands on 3+ platforms hold most mentions)
- Mention × citation quadrant reviewed: defend mentions, or push citability
- Reports include what changed, not promises: no citation guarantees, no model-update timelines
