Blog / What we measured
Who actually replies to cold LinkedIn outreach
A reply feels like progress. In September we looked at who was actually replying to our own outreach, and most of them could never buy what we sell.
The short version
- 265 people replied to our profiles between 15 and 26 September. About 25 of them, roughly 1 in 11, could buy a service like ours.
- 76 of them, nearly 3 in 10, sold lead generation, appointment setting or outreach themselves.
- The cause was our own searches. When you search LinkedIn for an activity, you find the people who sell that activity.
- We swapped 32 searches for ones aimed at the people who own a sales number, and taught our scoring to move sellers of outreach to the back of the queue.
Why we looked
LLM Leads runs LinkedIn outreach for B2B sales teams, and we run our own sales on the same system. Replies were coming in every day, and plenty of them turned into real conversations. Sign-ups were not following. So we stopped counting replies and started asking who was sending them.
How we sorted them
We took everyone who replied to one of our profiles between 15 and 26 September 2026: 265 people. For each one we kept only what their LinkedIn headline and company said about them, with no names. A model sorted each person into one of four groups, and we checked a sample of its calls. The groups were:
- Buyers. Founders, CEOs and sales leaders at a company that sells to other businesses and needs more sales conversations.
- Sellers of outreach. People whose own job is lead generation, appointment setting, cold email or LinkedIn outreach for clients.
- Individual contributors. SDRs, account executives, marketers, analysts and freelancers. Useful people, but not the ones who sign.
- Unclear. The headline did not say enough.
What we found
| Who replied | People | Share |
|---|---|---|
| Buyers | 25 | 9% |
| Sellers of outreach | 76 | 29% |
| Individual contributors | 155 | 58% |
| Unclear | 9 | 3% |
The sample check held up. A few of the buyer calls were generous (a marketing manager is not really the person who signs), so treat 9% as the ceiling rather than the floor.
The sellers stood out. They reply quickly and warmly, because a conversation with a sales team is exactly what they are looking for. Several of them answered our message by pitching us their own service.
The cause was our searches
Every profile finds its prospects through LinkedIn people searches. When we lined each reply up against the search that had found that person, the pattern was plain.
| Search | Replies | Buyers | Sellers of outreach |
|---|---|---|---|
| "linkedin outreach" | 13 | 0 | 8 |
| "sales navigator" | 11 | 0 | 5 |
| "cold email" | 5 | 0 | 4 |
| founder AND "lead generation" | 5 | 0 | 4 |
| "co-founder" AND SaaS | 6 | 3 | 2 |
| "pipeline generation" | 5 | 2 | 0 |
Each search found only a handful of repliers, so no single row proves much. Together they point one way. Searching for the activity (outreach, cold email, the tool people use to do it) finds the people who do that activity for a living. Searching for the person who owns the problem finds buyers.
Our scoring made it worse. It gave points for words like "lead generation", "outbound" and "clients", and a freelancer who sells outbound uses all of them in one headline. So the people least likely to buy were rated the best fit and contacted first.
What we changed
We removed 32 searches that were finding sellers of outreach and replaced them with 32 aimed at owners and sales leaders: founders, CEOs and co-founders of companies that sell to businesses, and heads of sales, sales directors and revenue leaders, across IT services, fintech, cybersecurity, logistics, HR software and funded startups.
We also changed the scoring. A headline that offers lead generation, appointment setting or outreach as a service now goes to the back of the queue. We did not remove those people entirely: some of them do buy, just not first.
The first run of the new searches, on one profile, found 98 people. The scoring rated 81 of them a strong or good fit. We read a sample: founders and CEOs of an IT services firm, a logistics network, a cybersecurity company, B2B marketplaces and AI startups. One seller of lead generation slipped through, with wording our scoring did not know yet, and it knows it now.
If you run outbound yourself
- Sort your replies by who sent them, not by how many there are. A reply rate can rise while the share of buyers falls.
- Search for the owner of the problem, not the activity. "Head of sales" at a software company is a buyer. "Outbound" in a headline is usually a competitor.
- Check what your scoring rewards. If it counts the vocabulary of your own industry, it will rank your competitors first.
We will check the same numbers again in a few weeks and publish how the mix of replies has moved.