LinkedIn InMail vs connection requests: which one actually gets replies for cold outreach
InMail reaches anyone but burns credits and often underperforms. Here is when connection requests win, when InMail is worth it, and how to decide.
Every LinkedIn outreach programme eventually asks the same question: do you send connection requests and message people once they accept, or do you pay for InMail and message them directly? The two paths look similar from the outside — a message lands in a prospect’s inbox either way — but they behave nothing alike on cost, deliverability and reply rate. Choosing the wrong one wastes either credits or weeks.
This is a comparison of the two channels as cold outreach tools: what each one costs, where each one wins, and how to decide which belongs in your sequence. It is written for teams sending at volume, not for one-off messages to a warm contact.
How the two channels actually differ
A connection request is an invitation. You send it — with or without a note — and nothing else happens until the person accepts. Once they do, you are first-degree connections and can message them freely, and those messages land in the primary LinkedIn inbox alongside notes from colleagues and people they know. The catch is the gate: no accept, no message. And LinkedIn caps how many requests an account can send, in the region of 100 to 200 per week for an established account, with tighter effective limits for newer accounts.
InMail skips the gate. It is a paid message you can send to anyone, connected or not, and it is metered by credits tied to your subscription. Sales Navigator Core includes 50 InMail credits a month; credits replenish monthly and roll over to a cap. Because InMail bypasses acceptance, it looks like the faster path. In practice it lands in a part of the inbox that prospects associate with recruiters and sales pitches, which is exactly the frame you do not want for a cold message.
What the reply rates tend to say
Published InMail response rates from LinkedIn’s own material sit in the region of 10 to 25%, and that is the number sales teams usually quote when they justify the spend. It is a real number, but it is measured across every kind of InMail, including recruiter outreach to active job seekers, who reply far more readily than a VP of Finance being sold software.
For cold B2B sales specifically, our own experience is that an accepted connection followed by a well-timed message tends to out-reply a cold InMail to the same person — often by a meaningful margin, though the gap varies by title and industry and we would not put a single figure on it. The reason is structural, not clever copywriting. When someone accepts your request, they have made a small choice to let you in. That choice primes them to read what comes next. An InMail arrives with no such choice behind it, so your first line is doing all the work of earning attention that an accept would otherwise have done for you.
The trade-off, then, is reach versus warmth. InMail buys you reach — you can message someone today without waiting for an accept, and without spending a request against your weekly cap. Connection requests buy you warmth — a smaller number of conversations, each starting from a prospect who already opted in. Most teams doing steady outbound get more pipeline per hour from warmth, which is why the connection path is the sensible default and InMail is the exception you reach for deliberately.
When InMail earns its price
There are cases where InMail is the right tool, and they are specific:
The account has hit its connection ceiling. If a seat is already sending its safe weekly maximum of requests, InMail adds reach without adding to that count, because the two limits are separate. It becomes a way to work a few high-value prospects the request budget cannot cover this week.
The prospect is hard to reach any other way. Some senior titles keep their connections tight and rarely accept from strangers. If the person is worth a direct, well-researched message and unlikely to accept a request, an InMail — treated as a one-shot, genuinely personalised note — can be the only door.
You are testing a segment before committing request volume. InMail can be a cheap probe: send a handful to a new title or industry, see whether the message resonates, and only then point your connection-request budget at the segment at scale.
What InMail is not is a volume channel. Fifty credits a month does not run an outbound programme, and buying more does not fix the underlying issue that a cold paid message starts colder than a warm accepted one. If you find yourself wanting hundreds of InMails a month, the honest read is that you need more connection-request capacity, not more credits. That is a question of seats and account health, and it is worth reading how VSDR runs LinkedIn outreach across dedicated seats rather than pushing a single account past its limits.
How to decide for your own programme
The decision is not really InMail or connection requests — it is where each fits in a sequence that leans on the free channel and spends the paid one carefully.
A practical default:
- Lead with connection requests to the whole target list, with a short or no note, and message on accept. This is your volume engine and it costs nothing but request budget.
- Reserve InMail for a shortlist: the prospects who did not accept within two or three weeks and are senior or strategic enough to justify a credit.
- Never send both at once to the same person in the same week. A pending request and an InMail arriving together reads as pressure, and pressure lowers reply rates on both.
- Track reply rate by channel and by segment, not as one blended number. The blended figure hides the fact that InMail may be carrying one hard-to-reach title while dragging on everything else.
The teams that get this right treat InMail as a scalpel and connection requests as the operating budget. If you want to see how that split maps to seats and monthly volume before you commit, the VSDR plans and what each seat includes lay out the connection-based model plainly, without an InMail dependency.
Where this comparison does not hold
Two situations break the default.
The first is recruiting and talent outreach, where InMail genuinely performs well. Active and passive candidates expect to hear from recruiters through that channel, and the reply rates reflect it. If your outreach is hiring rather than selling, the calculus flips and InMail may deserve to lead. This post is about B2B sales, where it usually should not.
The second is a brand-new or low-trust account. A fresh account with a thin profile and a low acceptance rate will struggle with both channels, but it will struggle with connection requests first, because low acceptance accelerates restrictions. In that narrow window, a small volume of carefully written InMails can keep conversations going while the account earns the history it needs. That is a bridge, not a strategy — the moment the account is warm enough to run requests safely, the economics point back to the free channel.
And a caveat that applies to both: neither channel rescues a weak list. An InMail to the wrong person and a connection request to the wrong person fail in the same way, one just costs a credit to do it. If your reply rates are poor across both channels, the problem is upstream in targeting, and no amount of channel tuning fixes it. Getting the list right is the work that makes either channel worth running.
