Flux
Guides

Published · Updated

Which Topics Attract Which Audience on LinkedIn

Edge

Your best-performing topic is not always your most valuable one. A post about salary negotiation can collect 600 reactions from job seekers while a quieter post about managing a first team reaches the HR leaders who might hire you. Engagement counts cannot tell those apart. The people behind the engagement can.

This guide shows how to compare the audiences of different topics with Flux's AI integration. You find your strongest topics, save each topic's posts as a collection, collect the engagers for each collection, and compare the groups by headline. The whole comparison usually costs under 25 credits.

The prompts on this page are things you type in Claude with Flux connected. Claude picks the tools; the tool names link to the MCP tool reference if you want the details. If you have not connected Flux yet, see Connecting to Claude.

What you'll learn

  • Finding your strongest topics and the top posts in each
  • Saving each topic's posts as a collection so the set can be analyzed together
  • Collecting the engagers for each topic, and what it costs
  • Comparing audiences by headline: job seekers, recruiters, peers, and buyers
  • Deciding when to pay for job history
  • What this method cannot tell you

Step 1: Which topics are worth comparing?

Start by asking which topics work for you:

Ask Claudefree

Which topics show up most in my posts, and which ones beat my average engagement?

Claude uses get_topic_insights. It ranks the terms and phrases that appear most often in your posts, and for each one it reports average engagement and vs_baseline, the topic's engagement relative to your own average.

Pick two to four topics that either beat the baseline or matter to your business. Then find the strongest posts in each:

Ask Claudefree

Find my two best-performing posts about salary negotiation and my two best about first-time managers.

Claude uses search_posts, a keyword and semantic search over your stored posts, or get_posts sorted by engagement when you want the top posts rather than the most relevant ones. Two posts per topic is enough for a first comparison. More posts give you a steadier picture, and the cost math in Step 3 tells you what each extra post adds.

Step 2: How do you group a topic's posts?

Save each topic's posts as a collection:

Ask Claudefree

Save the two salary negotiation posts as a collection called "Topic: salary negotiation", and the two first-time manager posts as "Topic: first-time manager".

Claude uses create_collection and add_to_collection. A collection keeps the set together, so you can rerun the comparison next month on the same posts. For a one-off comparison you can skip this step: Claude can collect engagers on a list of up to 50 posts directly.

Keep one rule in mind: do not count a post and its repost as two samples. A reshare of your own post usually draws many of the same people, so putting both in one collection double-counts that audience and makes the topic look more loyal than it is. Use the original post, or pick a different post on the same topic.

Step 3: How do you collect the engagers for each topic?

Ask Claudeabout 2 credits per post

Collect everyone who reacted to or commented on the posts in both topic collections.

Claude runs get_post_set_engagers once per collection. Each call returns one deduplicated list of the people who reacted, commented, or replied to comments on that collection's posts, ranked by how often they engaged. Every engager comes with their name and LinkedIn headline at no extra cost.

What it costs. Flux reads reactions and comments in pages of up to 100 people, and each page costs 1 credit. A good estimate is:

credits ≈ the number of reaction pages + the number of comment pages, summed across posts

A post with fewer than 100 reactions and fewer than 100 comments costs 2 credits. Posts collected in the last 30 days are skipped for free, so rerunning the comparison next month only pays for posts you add. By default Flux reads up to 10 pages of reactions and 10 pages of comments per post.

Here is the cost for a typical 4-post comparison, two posts per topic, each with fewer than 100 comments:

PostReactionsReaction pagesComment pagesCredits
Salary negotiation, post A586617
Salary negotiation, post B288314
First-time manager, post A333415
First-time manager, post B165213
Total1,37215419

About 19 credits buys you the names and headlines of roughly 1,370 people who reacted, plus everyone who commented. Your balance must cover the most a call could cost before it runs: 2 posts × 10 pages × 2 lists, or 40 credits per collection at the defaults. You are charged only for the pages actually fetched. If your balance is lower, Claude gets back smaller retries that fit, such as reading fewer pages per post.

Step 4: How do you compare the audiences?

Ask Claudefree

Sort each topic's engagers by headline into job seekers, recruiters, peers, and buyers. Compare the mix, how many engaged with both posts, and how much of each group's engagement is comments.

Headlines are self-written, but they are usually specific enough to place people into a handful of groups that matter for your business. For a career coach, those groups might be:

  • Job seekers: "Open to work", "Seeking my next role", "Former X, exploring opportunities"
  • Recruiters and talent teams: "Technical Recruiter", "Talent Acquisition Partner"
  • Peers: other coaches, consultants, and creators in your space
  • Buyers: the people who could hire you, such as HR leaders, people-team heads, and managers with a training budget

Then compare three things across topics.

The mix. What share of each topic's engagers falls into each group? This is the core answer to "which topic attracts which audience."

Regulars versus one-time engagers. Each engager carries an engagement_count and the list of posts they engaged with. People who engaged with both posts in a collection are that topic's regulars. A topic with many regulars has a loyal audience; a topic made almost entirely of one-time engagers is reaching new people who may not come back.

Comment share versus like share. Comments carry more intent than reactions. Compare what share of each group's engagement is comments. A group that comments heavily is in conversation with you; a group that only likes is watching. The post-level split between likes, comments, and reposts comes from each post's counts, which cover every post even before you collect engagers.

Step 5: When should you pay for job history?

Headlines answer "who is this person" for most comparisons. They do not reliably answer "which company" or "which industry." If those matter, look up job history only for the top engagers in each group rather than the whole list.

Ask Claude3 credits per person

Look up job history for the 15 most engaged people in the first-time manager group.

Claude uses enrich_people, or reruns get_post_set_engagers with a max_enrich limit. Either way it costs 3 credits per person found, and the posts themselves are not fetched again. People gain a current company, title, and full job history.

After that, get_post_engagers_by_company and get_engagement_industry_breakdown can group them. Both of those tools only cover engagers whose company is known, so everyone without job history lands in an unknown bucket. Read the results as the mix among your top engagers, not as your whole audience.

Worked example: a career coach compares two topics

Dana is a fictional career coach who sells two things: one-on-one coaching for individuals and manager-training workshops for companies. Topic insights show that "salary negotiation" runs well above her baseline, while "first-time manager" sits near it. She wants to know whether her strongest topic is also reaching workshop buyers.

She saves two posts per topic as collections and collects the engagers on both. The pull costs 19 credits, matching the table above.

Claude sorts the headlines and reports:

Salary negotiation (874 reactions)First-time manager (498 reactions)
Job seekersabout 45%about 10%
Recruiters and talent teamsabout 15%about 5%
Peers (coaches, consultants)about 20%about 25%
Buyers (HR and people leaders, managers)about 10%about 40%
Unclear headlineabout 10%about 20%
Engaged with both posts in the topic6%14%
Comments as a share of engagement3%7%

The salary-negotiation posts win on reach, and they build her audience of individual coaching clients. The first-time-manager posts reach a smaller crowd, but that crowd is mostly the people who approve workshop budgets, and they come back and comment more often.

Dana then looks up job history for the top 15 engagers in the first-time-manager group for 45 credits. Grouping each post's engagers by company shows three people from the same mid-size software company among them, which gives her a specific account to follow up with. Her content plan changes too: she keeps salary negotiation for reach and posts about new managers every week for pipeline.

What can't this tell you?

  • Impressions and viewers. LinkedIn does not expose who saw a post, so Flux cannot tell you about people who read without reacting. This method compares the people who engaged, not everyone who was reached.
  • Very large posts in full. Reaction and comment lists stop at 10 pages (about 1,000 people) per post by default. The result marks a post as truncated when more engagers were available, and Claude can raise the limit to 20 pages if you need them.
  • Company and industry for everyone. Those come only from job history, which Flux has for people you enrich or already looked up. Headlines are self-written and sometimes vague, so expect a share of every group to stay unclear.
  • Cause and effect. A topic's audience also depends on timing, format, and who reshared the post. Compare posts of a similar format and age when you can, and treat small samples as a direction rather than a verdict.

FAQ

How many posts do I need per topic?

Two posts per topic is enough to see a clear difference in audience mix. Use three to five if the topics are close or the posts are small.

Can I run this on someone else's profile?

Yes. On Edge, track the profile or company page, or pull its recent posts with linkedin_profile_posts, then collect the engagers on those posts the same way you do for your own. This is a useful way to see which topics bring a competitor its buyers. Collecting engagers costs credits the same way it does for your own posts.

Why does job history cost extra?

Job history comes from a separate profile lookup for each person, which returns their current company, title, and past positions. The headline that comes with every engager is free. Job history costs 3 credits per person.

Does rerunning the comparison cost the full amount again?

No. Posts collected in the last 30 days are skipped for free, so a rerun only pays for new posts. Ask Claude to force a refresh if you want to recollect posts inside that window.

On this page