| name | 51-audience-research-global |
| description | Use when paid ad AUDIENCES must be researched and defined before launch — target profile, interest and behavior mapping per platform, audience sizing, cold, warm, and hot tiering, lookalike seeds, exclusions, and ranked targeting hypotheses to test. Trigger on 'audience research', 'who should I target', 'interest targeting', 'Meta audience for this product', 'TikTok targeting ideas', 'my targeting is too broad'. Also use when launch is imminent and the ad set targeting is still empty. Not for — persona and JTBD depth for content, see `09-customer-insight-global`; warm retargeting tiers, see `56-retargeting-plan-global`; where the budget goes, see `54-media-plan-global`; the campaign hierarchy, see `52-account-structure-global`. |
| metadata | {"version":"1.0.1","category":"performance"} |
| license | MIT |
| triggers | ["audience research","target audience","interest targeting","audience for campaign","who should I target","Meta audience","TikTok audience","lookalike seed"] |
| output | File .md — audience profile, ready-to-paste targeting settings per platform, cold/warm/hot tiers, lookalike seed brief, and a ranked list of targeting hypotheses to test |
| related | ["product-marketing-context-global","09-customer-insight-global","08-competitor-research-global","10-reverse-kpi-global","52-account-structure-global","54-media-plan-global","56-retargeting-plan-global"] |
Audience Research (Global)
Wrong targeting burns budget no matter how good the copy or creative is. This is the first step of the performance chain: 51 -> 10-reverse-kpi-global -> 54-media-plan-global -> 53-tracking-setup-global -> 52-account-structure-global. If there is no customer insight yet, run 09-customer-insight-global first.
Information gathering
Read .agents/product-marketing-context-global.md and any output from 09-customer-insight-global. If information is missing, ask up to 4 questions:
- What product or service will run ads? Price point and market tier (mass / mid / premium)?
- What existing customer data is available? Age, gender, geography, purchase behavior, past customer list, CRM export, pixel data.
- Which platforms and which markets? Meta / Google / TikTok / YouTube / LinkedIn / Pinterest — and which countries. Primary objective: lead gen / conversion / traffic / awareness?
- Planned budget and target CPA/CPL? If not calculated yet, run
10-reverse-kpi-global.
Principles
- Geography is the biggest cost lever, before interests. Per
references/benchmarks-global.md, Tier 1 markets (US, Canada, Australia, Western EU) run 6-7x the CPM of Tier 2 (SEA, LATAM). US Meta CPM sits at $15-25 vs $2-6 in Brazil/LATAM. Decide the market before debating interest stacks.
- Research before copy and before campaign build. Never the reverse.
- Broad first, narrow only with evidence. Modern delivery algorithms optimize well on a wide pool plus strong creative. Narrow only when segment data proves it.
- One clear interest theme per ad set. Stacking many unrelated interests makes it impossible to tell which one worked.
- Write pain points in the customer's own words, not marketer language.
- Size for spend, not for a magic number. A cold ad set must be large enough to spend its daily budget for at least 7 days without frequency passing 2.5. A tight interest stack saturates fast in a Tier 1 market where CPM is $13-20.
- This is a living document. Update it after 3-5 days of live data by comparing CPA per segment.
Workflow
1. Core audience profile
Build from real data (CRM, analytics, order history, platform audience insights) plus 09-customer-insight-global:
