The people worth finding are not your competitor's customers, they are the strangers in their comment sections who have the problem and have not chosen yet. Watch the public conversations around a competitor, judge each person by what they wrote rather than where you found them, and never tell them where that was.
Disclosure: Signl is ours, and it runs the workflow below. Every product behaviour named here was checked against our code on 24 October 2026. The method needs no tool at all for one competitor and ten minutes a day.
Key facts
- Only public engagement is available: comments, reactions and reposts. Follower lists and connections are not, from any vendor.
- A competitor's thread contains at least three different people, and only one group is worth contacting.
- The fit test is what the person wrote, not that they appeared under a competitor's name.
- An opener that reveals where you found someone converts a warm signal into a creepy one. We think this is the fastest way to burn this channel.
- Four public surfaces exist around a competitor, and three of them are not the competitor's own posts.
Why a competitor's comment section is worth reading
Someone who comments under a competitor's post has done something a list cannot show you: they have put their problem into words, in public, next to a vendor in your category. That is a stage, a vocabulary and an objection in one place.
It is also the most sensitive source you can watch, for an obvious reason. The person was talking to someone else. Everything below exists to keep that from going wrong.
What it is not: a customer list. Treating it as one is the mistake behind most of the "steal your competitor's customers" advice on this query. Competitor threads skew heavily towards people who have not bought anything yet.
The four public surfaces
Most guides on this topic monitor a competitor's company page and stop. That is the thinnest of the four.
| Surface | What it finds | Why it is worth watching |
|---|---|---|
| Founder and executive profiles | The highest-engagement posts in most B2B companies | People reply to a person, not a brand. Volume and candour are both higher here |
| Company page posts | Product announcements, customers, launches | Lower engagement, but complaints and comparisons cluster under launches |
| Mentions of the competitor's name | People discussing them on someone else's post | Often the frankest material, and it is invisible if you only watch their page |
| Posts by their staff | Sales and marketing posting into your shared market | Their audience is your audience, assembled by someone else |
The third row is the one that repays setup effort. A mentions search finds the thread where someone says "we looked at Acme and it did not fit our stack", which will never appear under Acme's own post. In Signl a mentions search is labelled competitor by default, because that label is what switches the opener guard on; a search counts as 2 sources against the plan's 25, since it costs about twice as much to run as a profile.
Add their two or three most active people, not their whole company. A profile with no posts produces nothing and still consumes a source.
Judging fit from a comment
The question is not "did they engage with my competitor". It is "is this person someone I can help, and are they at a point where a message is welcome". Four tests, in order of how much they tell you. Run them before any contact lookup, so you are not paying to enrich the wrong group.
1. Did they write a problem, or a reaction? A reaction to the post ("agree", "this is spot on") tells you about the post. A problem ("our close takes nine days and reconciliation is still manual") tells you about them. Only the second is a reason to write.
2. Whose problem is it? "My team struggles with this" is a buyer. "Our clients struggle with this" is an agency or a consultant, which may be a partner rather than a prospect. "I used to see this" is a job seeker or a retiree. The pronoun is the fastest qualifier in this channel.
3. Are they already committed? Read for tense. "We're rolling out Acme next month" is closed. "We're looking at Acme" is open. "We left Acme last year" is open and holds an objection you can use. Treat the first as a follow in 6 months, not a lead.
4. Does the company fit at all? Size, country and industry still apply. A perfect comment from a company you cannot serve is not a lead, and this is where an ICP earns its keep. Signl checks each person against what you sell, the titles and seniorities you target, industries, countries and headcount bounds, and a known country outside your list caps the match whatever the rest of the comment suggests. Writing one that an automated check can use is covered in define your ICP.
The three people in every competitor thread
| Who they are | How they read | What to do |
|---|---|---|
| Current customers | Specifics about using it, feature requests, defending it | Leave them. Switching campaigns aimed at happy users waste the channel |
| Evaluators and the dissatisfied | Questions, comparisons, stated gaps, "does it do X" | This is the group. Contact them |
| Peers, vendors and job seekers | Praise, self-promotion, "open to work" headlines | Filter. Signl sets these aside free, before any model or lookup runs, with a reason recorded for each |
The second group is usually the smallest of the three, which is the point: the value of this source is concentration, not volume.
What an opener may say
Three rules. We enforce all three in code rather than in a prompt, because a prompt is a suggestion.
Never name the source. Not the competitor, not the post, not its author, not the fact that they commented. "Saw your comment on Acme's post about close" is the line that ends the conversation. Signl rejects a drafted opener on a competitor source that names the post's author or refers to their comment or post, and substitutes a plain line instead.
Quote only what is real. The fragment the opener answers has to appear verbatim in what they wrote. An invented quote in a cold message is worse than no quote. Signl checks the fragment against the comment and falls back to a template line when it does not match.
Answer the substance. Take the problem and respond to it as though they had said it to you, because in a sense they did, in public, about your category.
An illustrative example, not a customer story. Someone writes under a competitor's post: "We cut month end close to four days but reconciliation is still manual."
- Burns it: "Noticed you commenting on Acme's post about close times."
- Works: "Four days with reconciliation still manual is the gap I hear about most. The manual part is usually two or three specific reconciliations rather than the whole thing, in which case there is a shortcut worth ten minutes of your time."
Fifteen more rewrites, including the harder cases, are in how to message someone who commented on a post.
A public reply under their comment, before any message, is often the better first move. It is visible, it is proportionate to what they did, and it does not require an email address. Signl drafts one under 25 words with no pitch, no link and no product name.
Running it continuously
By hand: pick one competitor, add their founder's profile to a daily habit, read the new comments, act on two or three people. Fifteen minutes. Do this for two weeks before you automate anything, because it teaches you what your filter needs to catch.
With tooling, the job is to apply the same judgement every day across several competitors without contacting anyone twice. Signl watches profiles, posts, mentions searches and staff posts, filters the noise free, checks the rest against your ICP, looks up a verified work email only for the people who pass, and drafts the opener with the competitor guard on. Routing is a rule: people at or above a confidence, from these source types, into this campaign, at most this many a day, with a limit of 10 rules and 500 sends a day per account. The full seven-step version is LinkedIn comments to leads, and the tool landscape, including per-engager pricing from other vendors, is in LinkedIn engagement tracking tools.
Limits worth knowing before you build on this
- Private data is off the table. Followers, connections, post viewers and DM content are not public and not obtainable legitimately. Any offer of them should end the vendor conversation.
- Share-style post URLs do not expose comments. Links containing
-share-cannot be read for a comment thread. Watch the author's profile instead. - Large threads are sampled. Signl reads up to 500 people per post per check. Every tool has a ceiling; ask for the number.
- Automated collection is contractually contested. LinkedIn's User Agreement section 8.2 prohibits scraping and bot access (read 24 October 2026). That is a separate question from data protection law, which governs whether you may process and contact someone; our position on lawful basis is in GDPR and B2B prospecting. Neither question is answered by the comment being public. This is not legal advice.
- Objections are absolute. The UK ICO calls an objection to direct marketing an absolute right with no exemptions (read 24 October 2026). Signl suppresses and erases a person globally once they object, for every customer, permanently.
- This source rewards restraint. A competitor's audience is finite and it remembers. Three good messages a week from this channel is a better business than thirty.
Worth reading next: LinkedIn buying signals for the strength ladder behind test one, and LinkedIn social listening for where competitor sources sit among the others.
