Pick creators by who comments, not by follower count. Read the last ten posts, count how many commenters hold the titles you sell to, and watch for the same twenty names appearing every week. A creator with 4,000 followers and the right audience beats one with 90,000 and the wrong one.
Disclosure: Signl is ours, and it watches creator profiles as a source type. Product behaviour named here was checked against our code on 25 October 2026. The vetting method below needs nothing but a browser and twenty minutes.
Why a creator's comment section is a different source
A competitor's thread is full of people evaluating. A creator's thread is full of people thinking out loud, most of whom have not started looking yet. That changes both what you find and how you behave.
The practical difference is tone. On a competitor's post you are watching a purchase decision form. On a creator's post you are in a room where people are being candid about their work because the person who convened it has no product to sell them. Arriving with a pitch is more obviously out of place, and a useful contribution travels further.
It also changes the arithmetic. Creator threads are bigger and noisier, so the filter does more work and the hit rate per comment is lower. What you get in exchange is volume and repeatability: a good creator posts twice a week forever.
The ten-minute vetting scorecard
Most advice on this topic says to find creators your buyers follow, then stops. Here is how to actually check, using nothing but their recent posts.
Open the creator's last ten posts and run five checks. Each is a count, not a judgement.
| Check | What to count | Our working threshold |
|---|---|---|
| Buyer density | Of the commenters you can see, how many hold a title you sell to | One in six or better |
| Problem talk | How many comments describe a situation rather than praise the post | One in four or better |
| Repeat faces | How many of the same names appear across all ten posts | Under a third, or you are reading a pod |
| Vendor share | How many commenters sell to the same buyer you do | Under a quarter, or you are in a vendor echo chamber |
| Cadence | Posts per week, and whether comments arrive for more than a day | Twice a week, with a 48-hour tail |
Those thresholds are our view, not an industry standard. They are deliberately crude, because a creator that fails two of them clearly is not worth a week of your attention, and one that passes all five is worth watching for a quarter.
The first row is the one that decides it. Everything else is a reason to reject.
The red flag worth knowing
If the same twenty names appear under every post with on-topic but weightless comments, you are looking at an engagement pod rather than an audience. LinkedIn's Professional Community Policies tell members not to "agree with others ahead of time to like or re-share each other's content" (read 25 October 2026).
No tool detects this reliably, ours included. The manual test above is the detection method. Drop the source rather than trying to filter around it, because the signal you want is strangers being candid, and a pod is neither.
Finding the creators in the first place
Three methods, in increasing order of how much they tell you.
Ask five customers. Literally: whose posts do you read. This is the highest-quality answer available and almost nobody does it. Five answers will name two or three people repeatedly.
Read your own best-fit people's activity. For the prospects you already like, look at what they engage with publicly. You are not looking for the biggest name, you are looking for the name that appears twice.
Let the data say. After a fortnight of watching anything at all, the people who passed your ICP check have an engagement history. Signl suggests authors that two or more of your qualified people engaged with in the last 30 days, which is the second method run continuously instead of by hand; the rules are in sources Signl finds for you.
Do not start with a list of "top B2B voices". Those lists rank reach, and reach is the one variable that does not predict whether your buyers are in the room.
What to do with the people you find
The approach rules are stricter than on your own posts and looser than on a competitor's.
You may reference the topic. "The thread about month-end close this week had a lot of people saying the same thing" is fine, because the topic is public and the conversation was general.
You may not name the creator or the post. It tells the person they were watched, and it implies an association with someone who has not endorsed you. In Signl, a source labelled influencer allows the topic in a drafted opener and blocks the author's name; the relationship is set when you add the source and cannot be edited afterwards, because it governs what an opener is allowed to say.
Replying in the thread is usually better than a message. You are in a public room where being useful is visible to everyone else reading. A good reply under a creator's post is seen by more of your buyers than any message you could send, which is a genuine advantage this source has over the others.
The line-by-line version of what to write is in how to message someone who commented, and the strength of each kind of comment is in LinkedIn buying signals.
Running it continuously
By hand: one creator, read the comments on each new post, act on two or three people. Twenty minutes, twice a week.
With tooling, the job is the filter. A creator thread of 300 comments contains maybe fifteen people worth reading and three worth contacting, and finding those by eye every time is the part that stops happening by week three. Signl watches the profile, collects commenters and reactors, drops the empty praise and the job seekers free, checks the rest against your ICP, and only then looks up a verified work email, so the plan's 500 a month go to people who fit rather than to a thread. Very large posts are read up to 500 people per check.
The wider setup, including how creator sources sit alongside your own posts and competitors, is in LinkedIn social listening for B2B sales. If you would rather cause the conversation than watch someone else's, that is LinkedIn posts that bring buyers into the comments.
Limits
- You see commenters and reactors, nothing else. Follower lists, connections and post viewers are not public and not obtainable legitimately from any vendor.
- The best creator for your market may not exist. Plenty of niches have no commentator with a real audience. Watching a mediocre one because the method says to is worse than watching a competitor instead.
- Audiences drift. A creator who changes what they write about takes their comment section with them. Re-run the scorecard quarterly.
- You are borrowing someone else's room. Behave accordingly: the fastest way to lose this source is to make the creator's audience feel harvested, and creators do notice.
- Automated collection is contractually contested. LinkedIn's User Agreement section 8.2 prohibits scraping and bot access (read 25 October 2026), which is a separate question from whether data protection law lets you contact someone. Our position on lawful basis is in GDPR and B2B prospecting. This is not legal advice.
