The reply is not the meeting

LinkedIn gets answered, email reaches everyone, and phone turns contact into conversation: the best outbound channel changes every time you change the denominator

Ask which outbound channel works best and you can get three confident, contradictory answers before lunch. Email people point to the number of meetings email produces. LinkedIn people point to reply rate. Callers point to the conversations that would never have happened in an inbox.

The argument survives because everybody changes the denominator. A reply per message, a meeting per person, and a meeting per hour are different measurements. Put them in one leaderboard and the result looks precise while telling you almost nothing.

There is another reason to be careful. Most large outbound datasets are published by companies selling sales data, sequencing, or outsourced prospecting. They are useful, often unusually transparent, and not a census of B2B. The exact percentages below are benchmarks from the people who published them, not natural constants.

Email wins the race it was built to run

In a 2025 dataset of more than 7.5 million cold emails, Belkins found a 0.45% reply rate per email sent. That is far below the 3% to 10% figures passed around in sales posts, partly because Belkins used every sent email as the denominator instead of opens, delivered emails, or unique recipients.

That number makes email look broken. It also explains why email remains enormous. A bad per-send return can still produce the most total replies when sending another message costs almost nothing and can be automated. In a separate survey, 79.7% of salespeople said they use email for outbound, making it the most widely adopted channel in the sample.

The trade is simple: email buys cheap reach by making each individual touch easy to ignore. That is a reasonable trade for a team addressing a large market. It is a terrible excuse for a founder to send ten thousand messages before learning whether the first ten said anything worth answering.

LinkedIn wins replies, after an extra gate

Belkins and Expandi analyzed 15.1 million LinkedIn contacts from 2025 and reported an overall message reply rate of 7.2%. Already-connected Messenger campaigns reached 12.2%. Those numbers are startling next to 0.45% for email, and they are not directly comparable.

LinkedIn puts a gate before the message. The average connection acceptance rate in the larger dataset was about 26%, and the highest reply rate belonged to people already connected to the sender. In Belkins' smaller contact-level dataset, 1.3% of connected prospects eventually booked a meeting. A reply is progress, but it is not a meeting, and a reply from someone who first accepted you is not the same unit as an email sent to a stranger.

What LinkedIn buys is identity. Before deciding whether to answer, the recipient can see your face, role, company, history, and mutual connections. The message arrives with context an email has to earn in its first sentence. That makes LinkedIn especially strong where being on LinkedIn is already part of the job: recruiting, professional services, venture capital, and much of B2B software.

Phone loses the first nine seconds and wins the next ten minutes

The phone data looks brutal from the top of the funnel. Across more than 175,000 dials, Belkins found that 9.9% reached a person, 58% of those connects became actual conversations, and 4.6% of conversations became meetings. End to end, that was roughly one meeting per 370 dials.

Look at the bottom of the same funnel and phone changes character. One meeting per 370 dials is poor. One meeting per 22 genuine conversations is not. The expensive part is getting a person into the conversation; once they are there, the channel collapses several email exchanges into a few minutes. Questions get answered, objections surface, and the next step can be agreed before either person opens another tab.

Two surveys disagree because they asked different people

A broad 2025 sales survey asked which channel was most effective at eliciting a cold-outreach response. The answer was LinkedIn or social at 36%, email at 25%, and phone at 16%. A 2026 survey of BDRs reported almost the reverse for outbound: phone at 40%, email at 36%, and social at 10%.

This is less contradiction than it looks. The first survey asks a broad population about responses. The second asks people whose job is to create qualified sales conversations. LinkedIn is very good at making a stranger answer. Phone is very good once the answer itself is the beginning of the sales conversation.

The sequence gets the credit, then the last touch takes it

Modern outbound makes channel attribution even less honest, because the channels are not being used alone. Salesloft's analysis of 3.4 million cadences found that 80% of its top 100 cadences began with a call followed by an email. In Belkins' follow-up programs, calling accounted for 33.6% of appointments even though the call often came after unanswered email.

Which channel earned that meeting? The call created the conversation. The earlier email may have made the name familiar enough for the call to be answered. LinkedIn may have shown the recipient that the caller was real. Giving the entire meeting to the last touch is convenient for a dashboard and wrong about how recognition accumulates.

Start where the buyer already works

The right opening channel is mostly a fact about the buyer. A recruiter, consultant, or software founder is professionally visible on LinkedIn, so starting there can carry useful context. A plant manager, restaurant owner, or construction contractor may be much easier to reach by phone. A civil servant or university administrator is likely to have a public, stable work email and a process that prefers it.

Geography moves the baseline too. In the LinkedIn study, the Netherlands produced a 10.5% reply rate against 6.0% in the US. That does not mean Dutch buyers love prospecting messages. It means a US benchmark can be the wrong prior for a European campaign, especially when saturation and professional habits differ.

A founder sequence small enough to learn from

1. Find the reason now. Start from a dated, public sales trigger, not a list of companies you already decided to bother.
2. Use LinkedIn for context. Check that you have the right person and learn how they describe the work. Connect when the platform is natural for their role.
3. Use email for the complete note. One relevant fact, one credible connection to the problem you solve, one small ask.
4. Use phone to resolve, not repeat. The call should advance the thought in the email, not read it aloud.
5. Measure people through the funnel. Track prospects reached, genuine conversations, meetings, and what each prior touch contributed. Do not celebrate a reply leaderboard.

Run that across ten accounts a week and you will learn more than a thousand-send benchmark can tell you about your own market. If LinkedIn creates conversations, lean into it. If the buyer lives on the phone, call earlier. If email keeps producing meetings cheaply, let it carry more of the load. The sequence is a hypothesis, not a ritual.

The channel is not the reason someone answers. It is the route the reason travels. OctoLoops finds companies with a current, sourced trigger and drafts the first note from that fact, so you can spend your time deciding how each person should be reached. If you want to try the loop on your own product, the plan is free.