audience-discovery
Skill1 starsChanged 57 days agoMITView on GitHub
Use when stage 4 of the prospecting system needs the places where buyers like your best customers already gather — competitor pages, adjacent tool pages, communities, and events — found, sized, and ranked before any audience is bought or scraped. Also use when the ask is \"whose followers should we target?\".
In this repository21 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- account-notesSkill
- audience-filterSkill
- buying-signalsSkill
- cold-emailSkill
- contact-enrichmentSkill
- customer-patternsSkill
- follower-scrapeSkill
- follow-upsSkill
- icp-definitionSkill
- lead-prioritizationSkill
- lead-qualificationSkill
- lead-scoringSkill
- linkedin-outreachSkill
- lookalike-researchSkill
- personalizationSkill
- pipeline-reviewSkill
- prospecting-systemSkill
- reply-classificationSkill
- segmentationSkill
- targeted-searchSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 35
Rank: Prefer pages that are narrow and on-problem over pages that are big.
- Distinctive: line 50
Do not buy anything in this stage.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Audience Discovery
Your next customers already follow someone. This skill finds who, and how much it costs to reach that audience, so stage 5 only buys the ones worth buying.
Pipeline contract
Stage 4 of 20. Reads 01-customer-patterns.md, 02-client-profile.yaml. Writes 04-audiences.csv.
Files live in runs/<run>/. Column definitions: /prospecting-system.
Four source types
audience_type | What to look for | How people get collected in stage 5 |
competitor | Direct competitors and the vendor your customers replaced | LinkedIn page followers; post engagers |
tool | Software your customers already use that signals the problem exists (from the tech/stack findings in stage 1) | LinkedIn page followers |
community | Slack/Discord groups, subreddits, newsletters, LinkedIn groups your customers' champions are in | member export or CSV import; Reddit authors |
event | Conferences, webinars, meetups your customers attend or sponsor | event LinkedIn page followers; attendee CSV |
Steps
- Start from evidence, not brainstorming. For each customer in
00-customers.csv: which competitor did they replace or evaluate, which tools are on their site and job posts, which events do they sponsor or speak at, which communities do their champions mention on LinkedIn. One row of evidence per find. - Expand. For each evidenced page, find 2–3 siblings (the competitor's competitors, the tool's category peers). Mark these
evidence=inferred. - Resolve the LinkedIn company page URL for every competitor, tool, and event row. No page found = leave
linkedin_company_urlblank and keep the row. - Size and price each page with GetLeads MCP
lookup_company_linkedin_followers(free lookup; pass the profile's headcount, seniority, and country filters). Recordfollower_countandest_cost_usd($0.0035 per follower, $8.75 minimum per page — confirm against the lookup result, prices change). - Rank. Prefer pages that are narrow and on-problem over pages that are big. A 4,000-follower niche competitor beats a 900,000-follower platform: the big page costs more and the judge will reject most of it.
- Present the table and let the human pick. Show total cost of the picks against the run's remaining spend cap. Set
selected=trueonly on what they chose.
Output: 04-audiences.csv
audience_type, name, url, linkedin_company_url, evidence, evidence_detail, follower_count, est_cost_usd, collection_method, selected
evidence—customer(seen at a named customer) orinferred.collection_method—getleads_followers,post_engagers,csv_import, orreddit.
Gotchas
- A customer's own page is not an audience — its followers are their customers, not yours.
- Skip pages whose followers are mostly job seekers (large employers, recruiters' favourites) unless the judge sample proves otherwise.
- Communities rarely allow member export. Record them anyway: they are where stage 9 finds signals and where stage 13 finds language.
- Do not buy anything in this stage.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.