Most LinkedIn prospecting starts the same way: a Sales Navigator filter, a list of people who match a title and a headcount band, and a message that could have gone to any of them. It works, slowly, because some fraction of any large list is in the market this month.
There is a second way to do it that starts from the opposite end. Instead of asking who fits, it asks who just did something, then checks whether the people who did fit. It produces far fewer names and a much higher proportion of replies, and it is limited by a hard ceiling on how many people one LinkedIn account can contact at all.
This page covers both, the arithmetic that constrains them, and what the difference does to reply rates.
What is LinkedIn prospecting?
LinkedIn prospecting is finding people on LinkedIn who might buy what you sell, and starting a conversation with them. In practice it splits into two jobs that are often confused: sourcing, which is deciding who to contact, and outreach, which is what you say and when you say it.
Sourcing on LinkedIn happens one of two ways. Attribute-based sourcing filters the member base by title, company size, industry and geography, which is what Sales Navigator is built for. Activity-based sourcing starts from something a person did in public, such as a comment, a reaction, a post or a follow, and works backwards to whether they fit. Most tools do the first. The second is harder to build and produces a shorter list.
Is LinkedIn a good prospecting tool?
For B2B, yes, with one structural caveat: LinkedIn caps how many people you can reach, so the quality of the list matters more than on any other channel.
Most accounts are limited to roughly 800 connection invitations a month, and a healthy sending pattern runs 30 to 40 invitations a day once an account is warmed up. That ceiling does not move much with effort or spend. It means a LinkedIn account is a fixed-size asset: about 800 chances a month to reach someone new.
On email, a bad list costs deliverability and time. On LinkedIn, a bad list costs the ceiling itself, and the ceiling is the whole resource. That is the argument for spending effort on who rather than on how many.
How many people can you actually reach in a month?
One account, roughly 800 invitations a month, ramping up over the first three weeks. The ramp exists because a new account with no sending history draws attention faster than an established one.
| Week | Invitations a day | Why |
|---|---|---|
| 1 | 15 to 20 | The account has no history of sending anything |
| 2 | 25 to 30 | Acceptances start to establish a pattern |
| 3 onward | 30 to 40 | Steady, as long as acceptance holds |
Three numbers say whether an account is healthy, and none of them is messages sent:
- Acceptance rate. Below 20% means the people receiving invitations do not recognise why. That is a targeting problem, not a volume one.
- Pending invitations. Unanswered invites count against the ceiling. Withdrawing anything older than two weeks frees it up.
- Reply sentiment. One "who are you?" is noise. Several in a week means the opening line is wrong.
Where do the best prospects actually come from?
From people who did something recently, in public, that says what is on their mind. WarmDogo's production runs rank the sources, and the spread between best and worst is larger than most prospecting advice suggests.
A signal worth acting on has four parts: an actor, an action, a topic and a date. A filter gives you an actor and nothing else.
| Source | What it is | Yield |
|---|---|---|
| Voices your buyers follow | People who engaged with a post by someone your market pays attention to | Highest per attempt, and the most consistent |
| Phrase engagers | People who engaged with a post using a phrase that names the problem | Highest volume of qualified leads |
| Phrase authors | People who wrote a post using that phrase | Strong, lower volume |
| Hiring | Companies hiring for the role that owns the problem | Occasional, and needs a second step to find the buyer |
| Competitor followers | People engaging with a competitor's page | Lowest. Measurably the worst source in every run so far |
The last row is the one that surprises people, because competitor-audience targeting is what most tools advertise. Across one recent run, competitor pages produced a single lead from 24 attempts, and the leads it did produce were rejected on review. Voices, over the same run, produced eight leads from six attempts, six of which were kept.
The reason is selection. Someone following a competitor may be a customer, a job seeker, an employee or a rival. Someone who commented on a post about the problem you solve has told you something narrower and more recent.
What is the 3/2/1 rule, and what are the 5 P's?
Both come up constantly and neither is a LinkedIn policy. They are sales frameworks that circulate under several different definitions, and the versions disagree with each other. Treating either as a rule the platform enforces will not help.
What LinkedIn does enforce is the invitation ceiling, the weekly invitation limit, and its response to accounts whose invitations go unaccepted. Those are the numbers worth learning, and they are in the table above.
What to say first
If the signal has all four parts, the first line writes itself, because it refers to something the person actually did. If it does not, the message ends up describing them in general terms, which is the pattern every LinkedIn user has learned to skim past.
Two things decide whether the opener lands:
- Recency. A reference to something someone did on Tuesday reads as attention. The same reference a month later reads as surveillance. Most signals stop being useful within two weeks.
- Specificity. Name the thing. "Saw your comment on the thread about onboarding drop-off" works. "Saw you're interested in growth" does not, because it could be sent to anyone.
What is the best AI tool for LinkedIn prospecting?
It depends which of the two jobs you need done, and most tools do one.
| The tool does | What to check before buying |
|---|---|
| Sourcing only | Where the data comes from, and whether it is disclosed. You will still need something that sends |
| Sending only | Whether it respects the daily and monthly ceilings automatically, or leaves the pacing to you |
| Both | Whether you can see and approve what goes out before it goes, and whether sourcing goes beyond a filter |
Three questions worth asking whichever you choose:
- Does it show its reasoning? A lead that arrives without the reason it qualified cannot be reviewed, only trusted.
- Does it meter the thing you do most? Credit systems that charge per message make the cost of using the product unpredictable.
- Does it stop? A tool that sends the same volume on a day when nobody did anything interesting is spending the account's ceiling on noise.
WarmDogo does both jobs and prices per connected LinkedIn account. See pricing.
Where to start
Pick one voice your buyers actually follow and read who engages with their posts for two weeks. That single change produces a better list than any filter, and it costs nothing to test. If the replies improve, the rest of this page is about doing the same thing at the scale one account allows.