Fit and reachability are two different questions
Most LinkedIn targeting stops at fit. Does this person match the ideal customer profile? Right industry, right title, right company size. That question matters, and it is the one lead scoring tools like our own PipelineIQ were built to answer.
But a perfect-fit prospect who never opens LinkedIn is worth nothing to a connection campaign. There is a second question hiding behind the first one: will this person actually see the request and respond?
We call that second question reachability. It is not the same as fit, and it should not be blended into a fit score. A VP of Operations can be an ideal buyer and a terrible LinkedIn target at the same time. Mixing the two into one number hides exactly the thing you need to see.
So we went looking for evidence. Given only what is visible on a public LinkedIn profile before you send a request, can you predict who will accept?
What we looked at
The data came from real LinkedIn outreach history across five client programs. For each prospect we knew whether they had accepted the connection request. For about 2,300 of them we also had public profile signals and their posting activity over the previous 90 days.
We tested every signal we could see on a profile:
- posting activity in the last 90 days
- connection count
- whether there was a profile photo
- LinkedIn Premium
- creator mode
- whether the About section was filled in
- whether the current role was listed
- open-to-work status
- account age
The study compared people who accepted (with at least 30 days for the request to mature) against a random sample of people who did not, balanced within each program so one large campaign could not dominate the result. Every sample used a fixed seed, so the analysis can be rerun and produces the same answer.
Recent posting was the strongest signal
Two signals stood out, and they held up even after controlling for everything else: recent posting activity and connection count. After adjusting for the other signals, they were associated with up to roughly four to five and a half times higher odds of acceptance.
That led to three simple tiers anyone can apply by looking at a profile:
| Tier | What it looks like | Accepted |
|---|---|---|
| A: Recent posters | Posted on LinkedIn in the last 90 days | 27.4% |
| B: Active, not posting | Has a photo and 50+ connections, but no recent posts | 12.4% |
| C: Likely dormant | No photo or under 50 connections, and no recent posts | 4.9% |
A recent poster was more than five times as likely to accept as a dormant-looking profile.
One more flag was worth tracking on its own. Profiles with fewer than 100 connections accepted at just 1.9%.
A combined score ranks prospects well
Next we combined all the signals into one reachability model and tested it on prospects it had not seen, using five-fold cross-validation. It scored an AUC of 0.76. In plain terms, if you pick one person who accepted and one who did not, the model ranks the accepter higher about three times out of four.
The practical version is more useful. The top 20% of prospects, ranked by the model, captured 47% of all acceptances. That is more than double what random ordering would give you.
Replies are much harder to predict
Acceptance is only the first step. We ran the same analysis on replies after a connection was accepted, and the signals were much weaker. Posting activity tells you someone is present on LinkedIn. It does not tell you they want to talk about your offer.
That fits how outbound actually works. Getting the connection is mostly about presence. Getting the reply is about relevance, which comes back to targeting, the offer, and the message. Reachability gets you in front of the right person. It does not do the selling for you.
What this means for your LinkedIn outreach
Check for recent activity before you send. It takes seconds to see whether someone has posted in the last three months. With weekly invitation limits on LinkedIn, every request spent on a dormant profile is a request not spent on someone who will see it.
Prioritize, do not purge. The right move is to send to your Tier A prospects first, not to delete everyone else. Dormant-looking prospects still accept sometimes, and some of them are your best-fit buyers. Reach them another way, such as email, or save them for a later wave.
Keep fit and reachability separate. Score fit first, because an unreachable perfect-fit prospect is still a better long-term target than a reachable wrong-fit one. Then use reachability to decide the order and the channel. Two scores, two decisions.
Do the math on your own list. If these rates held for your market, 100 requests to Tier A prospects would produce around 27 new connections. The same 100 requests to Tier C would produce around 5. The list is the same size and costs the same effort. The result is five times different.
What this data cannot tell you
We want to be clear about the limits, because a number without its caveats is how bad advice spreads.
- It is observational, not causal. Posting does not make someone accept. It is a sign that they are active on the platform.
- Connection count is partly circular. People with many connections are often people who accept a lot of requests. That makes the signal useful for ranking, but it is a habit, not something your outreach changes.
- Profiles were captured after the outreach, not before. Going forward, the fix is to capture the signals before sending.
- It covers five B2B programs over one period. The real test is a new campaign where nobody is filtered out in advance, which is how we are validating it next.
- The post history was capped at five recent posts per person, and "dormant" is a heuristic, not a confirmed last login.
None of this changes the practical takeaway. It just keeps the claim honest: recent activity is a strong, cheap signal for ordering LinkedIn outreach, and it is not a guarantee.
If you want to know how your own targeting holds up, get a free Outbound Review. It takes about a minute, and it shows where your targeting is loose and what to fix first.
Frequently Asked Questions
What is a good LinkedIn connection acceptance rate?
It depends heavily on who you target. In our data, prospects who had posted in the last 90 days accepted at 27.4%, active profiles without recent posts at 12.4%, and dormant-looking profiles at 4.9%. Your overall rate is largely a function of the mix of those groups in your list.
What is the best predictor of whether someone will accept a LinkedIn connection request?
In our study of about 2,300 B2B prospects, recent posting activity and connection count were the strongest predictors. After adjusting for other profile signals, both were associated with roughly four to five and a half times higher odds of acceptance.
Should I remove inactive LinkedIn profiles from my outreach list?
No. Prioritize active profiles first, and reach inactive ones through another channel, such as email, or in a later wave. Some of your best-fit buyers are rarely on LinkedIn, so filtering them out entirely throws away good prospects.
Does LinkedIn activity predict replies too?
Much less. Profile signals predicted acceptance well but replies only weakly. Replies depend mostly on relevance: the right person, the right offer, and a message worth answering.
Who ran this study?
Leads By Me, a fully managed LinkedIn and cold email outbound agency based in Burlington, Ontario. The analysis used aggregate outreach history from our own client programs. No individual prospect data is published.