Community Intent Signals
Outcome: a small, high-intent list built from people describing your problem in public, with a message that references what they actually said.
- Surface
- App and MCP server
- Level
- Intermediate
- Uses
- linkedin_profile_posts · x_profile_posts · linkedin_post_commenters · find_work_email
- Credits
- ~20 for 15 contacts
- Prerequisite
- Lessons 05 and 06
The strongest signal there is
Someone writing “we’re evaluating options for X” or “our current setup for Y is falling apart” has told you three things at once: they have the problem, it is current, and they are willing to talk about it.
No firmographic comes close. The catch is volume — this produces tens of leads, not thousands — and etiquette, which decides whether the approach lands as helpful or as surveillance.
Where to look
| Source | What you find | Tool |
|---|---|---|
| LinkedIn posts and comments | Professionals describing work problems | linkedin_profile_posts, linkedin_post_commenters |
| X / Twitter | Faster, more candid, more technical | x_profile_posts, x_post_commenters |
| Competitor page engagement | People engaging with a competitor’s content | linkedin_page_posts |
| Job posts describing the problem | The company version of the same signal | company_job_listings |
The last row is worth noticing: a job post is a community signal written by the company rather than the person, and it is usually easier to find at volume. See Hiring Signals.
What counts as intent
Sort what you find into three tiers.
| Tier | Example | Action |
|---|---|---|
| High | ”We’re looking at tools for X — recommendations?” | Contact now, reference directly |
| Medium | ”Our X process is a mess right now” | Contact, reference the frustration not the solution |
| Low | Reacted to a post about X | Topic as hook, not the engagement |
Anything below that — following a competitor, being in an industry group — is not intent, it is adjacency.
From post to prospect
Capture the post and the person
Store the post URL, the text, and the date. The date matters: intent decays in days, not months.
Check the company qualifies
Free filter before any spend. An intent signal at a company you cannot serve is not a lead.
Check they are not a peer
Consultants, vendors and competitors ask these questions constantly.
Enrich and verify
Only survivors. Standard waterfall.
Write from what they said
Quote it.
The etiquette
This is where the play succeeds or backfires.
Do:
- Reference the specific thing they wrote, and quote it accurately.
- Be useful first. If they asked for recommendations, give a genuine answer — even one that is not you.
- Be direct about who you are. “I work on X” is fine; pretending to be a peer is not.
- Move fast. A week-old question has been answered by someone else.
Do not:
- Pitch in a public reply. Answer publicly if you have something useful, then follow up privately.
- Reference a reaction as though it were a statement.
- Contact someone who asked in a private or closed community. If it was not public, it is not a lead.
- Send the same template to everyone who used the keyword.
This play is small-volume by nature and stops working the moment it is automated at scale. Fifteen genuinely relevant, well-written messages a week will outperform 500 keyword-matched ones — and the 500 will get you named in the exact kind of post you are monitoring.
Do this now
Pick three sources
Two competitor or industry-voice pages, one keyword or topic you can monitor.
Pull recent posts and engagement
Last 14 days only.
Sort into intent tiers
High, medium, low.
Filter on company and role
Free. Drop peers, vendors and consultants.
Enrich the high and medium tiers only
Write individually for the high tier
Genuinely individually. There will not be many.
Template the medium tier lightly
Quoting their words, from the evidence column.
Measure reply rate against your cold baseline
If it is not several times higher, the filtering is too loose.
Check your work
- Every row has a post URL, text and date
- Only public sources were used
- Peers, vendors and competitors are removed
- High-tier messages are written individually
- Reply rate is compared against your cold baseline
Where this breaks
Keyword-matching without reading the post produces embarrassing outreach. Someone writing “we just finished migrating off X and it’s been great” matches the same keywords as someone struggling with X, and a message assuming they need help arrives as evidence that nobody read what they wrote. Read every high-tier post yourself. There are few enough that you can.
Further automation
Automate the monitoring, not the sending. A scheduled run that surfaces qualifying posts into a queue each morning gives you the speed advantage without the volume risk — a human still writes the message, but they start from a filtered list rather than a search box.
Next lesson
12 — Tech-stack fit lists, the last sourcing route in the course.
Reference for this lesson: linkedin_profile_posts, x_profile_posts, LinkedIn engagement, Signals & ABM.