LinkedIn or email for cold outreach: what each touch costs and returns
By Jānis Plūme, Founder, Outbound Pros · 2026-08-06
Quick answer
LinkedIn returns more per touch and email delivers far more touches, so the useful comparison is not which channel is better but what each touch costs and returns. On one client programme in professional services we measured 59% connection request acceptance, roughly 9% reply rate on LinkedIn DMs, and roughly 1.5% email reply rate, on the same accounts over the same period. LinkedIn replied at roughly six times email's rate on a fraction of the volume. That makes LinkedIn an escalation channel and email an opener on most lists. The reason is arithmetic, not preference.
Running email and LinkedIn as one sequence for B2B clients since 2024.
Is LinkedIn or email better for cold outreach?
Neither channel is better in general, because "better" collapses two separate quantities: response per touch and touches available. LinkedIn wins the first by a distance and loses the second by orders of magnitude, so the answer depends on how many people you need to reach and how long you have to reach them.
The answer is per list, and it comes out of one division.
What does each channel actually cost and return per touch?
Touch economics is the practice of comparing channels on what a single touch costs to place and what that touch returns, rather than on aggregate channel performance. It is the only comparison that survives a change in list size.
| Cold email | LinkedIn connection request | LinkedIn DM | |
|---|---|---|---|
| What caps the volume | Mailboxes provisioned and warmed | A hard per seat daily ceiling | A hard per seat daily ceiling |
| Scales with spend | Yes, by adding warmed mailboxes | No. More software does not raise a seat's ceiling | No |
| Lead time to add capacity | 4 to 6 weeks of warm up | Time to warm an operator on a new account | Same |
| Response per touch, same accounts | Roughly 1.5% reply rate | 59% acceptance | Roughly 9% reply rate |
| Cost of one more touch | Near zero | High. The ceiling is fixed and every request is spent | High |
| Position in a sequence | Opener on most lists | Escalation on the signal slice | Escalation after acceptance |
A channel that returns several times more per touch and delivers a fraction of the volume is a precision instrument and not an opener, and pointing it at everyone spends it on the list that was never going to engage. Email's scaling is real and it is not instant either: every mailbox in our sending estate had to be provisioned, authenticated and warmed before it carried anything, and the size of that estate is recorded once, with the rest of the group facts, on the page where those numbers live.
What do the reply rates look like on identical accounts?
The strongest comparison available is one where both channels ran against the same people, from the same accounts, in the same period. On one client programme in professional services we measured 59% connection request acceptance, roughly 9% reply rate on LinkedIn DMs, and roughly 1.5% email reply rate. That dataset belongs to our sibling property, published in full with its methodology at linkedpros.io on LinkedIn reply rates.
Same accounts, same period is the part that matters. Almost every published channel comparison sets one company's LinkedIn beside another company's email, with different lists, offers and senders. That is not a comparison, it is two anecdotes in a table.
Acceptance and reply are different events, so the 59% does not sit on the same axis as the reply rates. The comparison carrying the argument is 9% against 1.5%, roughly six to one.
Why does the scarce channel never open a large sequence?
The opening channel has to reach the whole list inside your window, and above a few thousand names LinkedIn cannot. The test is one division, the coverage window that governs every cadence we build.
coverage window (working days) = prospect count ÷ daily ceiling on that channelA 12,000 prospect list against a single LinkedIn seat produces a coverage window measured in months. A sequence built to run in three weeks cannot open on a channel that takes months to make first contact, so LinkedIn moves to escalation and email opens. That division settles the debate for the list in front of you, which is why a general answer does not exist.
The ceilings themselves are a research subject of their own, maintained by our sibling site in the connection request ceiling research. We treat its number as a fixed input. Full method on the capacity planning page.
When is LinkedIn genuinely the right opener?
LinkedIn should open when your seat capacity can make first contact with the entire list inside your window, which in practice means small, high value, signal triggered lists. A profile view and an accepted connection make the email that follows read as a second contact rather than a first. That advantage is only available when the arithmetic allows it.
We call that motion Spearhead. It runs alongside WideNET, our high volume angle testing across the full addressable market, which opens on email and escalates to LinkedIn on the slice that showed a signal. Spearhead runs on the hottest few hundred accounts and opens on LinkedIn.
The part nobody writes about is the approval gate. Every client approves the messaging and the lead lists before anything sends, one round, and the Spearhead slice is where that review gets real. Nobody meaningfully reviews a 30,000 record WideNET list. Clients argue in detail about 300 named accounts, and they should, because the seat capacity behind those 300 is the scarcest resource in the programme.
Why do published answers to this question disagree so badly?
Almost every published answer is a capacity artefact of the author's own list, presented as a property of the channel. The founder with a few hundred accounts correctly concludes LinkedIn. The agency running large volume correctly concludes email. Both are describing their own division.
Two things make it worse. A large share of the answers come from a vendor whose product is one of the two channels. And the question gets asked as though the channels were substitutes, when in our own fleet the sequences running both measured better than the single channel motion we publish from the same snapshot.
Is cold email dead?
No, and the claim usually rests on a metric error rather than on data. Comparisons in this category routinely place a ratio measured against replies beside a rate measured against sends. Those are different quantities, they differ by roughly two orders of magnitude, and setting them side by side manufactures a gap the underlying numbers do not support. Check whether both sides of any channel comparison share a denominator first. Our sibling property allboundpros.io takes that distinction apart properly once it publishes.
What did change is the floor. Mailbox providers tightened their bulk sender requirements, and Google publishes what it enforces in its own email sender guidelines: authentication, one click unsubscribe, a spam complaint ceiling. That killed the version of cold email that ran unauthenticated at volume and made deliverability a standing discipline rather than a setup task. That work sits with the email side of the group's agency.
What does running both actually get you?
In the Outbound Pros fleet snapshot dated 2026-07-06, sequences run as a coordinated multichannel motion produced a 0.37% positive reply rate across 8,714 emails sent, 7.36x the fleet baseline, against 0.14% on 52,786 sends for the follower sourced motion in the same snapshot. That is an observation from one fleet over one window and not a controlled test, and the denominator flatters multichannel, because positive replies arriving on LinkedIn are counted against emails sent. We publish that confound at the same size as the claim, together with the one sequence in the fleet that beat the multichannel segment average, on the benchmark page with the full working.
Which channel should you start with if you can only run one?
Start with the channel whose capacity matches your list inside your window, and where that leaves a choice, the one you can staff.
| Your situation | Start with | Why |
|---|---|---|
| Under roughly 500 named accounts | Seat capacity serves the whole list inside a normal window, and acceptance warms everything after it | |
| A few thousand to tens of thousands of prospects | LinkedIn cannot make first contact with that list inside any sane window | |
| No warmed mailboxes, you need signal this month | LinkedIn, on a small slice | Warm up does not compress. 4 to 6 weeks is 4 to 6 weeks |
| One seat, one operator, a list you cannot segment | Email, LinkedIn out of scope | A LinkedIn channel you cannot serve is a liability rather than a channel |
| Founder led, your own network, high ACV | LinkedIn, from your own profile | Your voice from your own account beats almost any agency copy in the first hundred touches |
The third row is the one clients find hardest to accept and the one I have most direct experience of. Onboarding runs about 21 days and mailbox warm up runs weeks on top of that, so a client who signs in the first week of a quarter has no real email volume until the back half of it. For those weeks the honest answer is LinkedIn, because email does not yet exist at usable scale. That is a fact about your build and not about the channels, and confusing the two is how teams conclude LinkedIn wins on six weeks of enforced patience.
What this comparison cannot settle
It cannot tell you which channel produced the meeting. There is no join key between a LinkedIn profile impression and an email open, so channel of reply is description and not causation, and a dashboard presenting it as attribution is reporting last touch. What is measurable is covered on attribution inside a sequence.
It cannot tell you your reply rates, which depend on the offer, the market and the list. The figures here come from one fleet and one client programme, not the category, and the sample and window caveats are published separately.
It does not cover compliance, where the channels are not equivalent. US cold email sits under a statute with named obligations, set out in the FTC's own CAN-SPAM compliance guide, with GDPR and national rules in Europe. LinkedIn messaging sits under LinkedIn's platform terms, a different risk and not a smaller one.
And it is the wrong question for two groups. If a sequence sits below the 0.5% positive on sends line where we kill rather than iterate, the channel is not the problem, and I have never seen a channel switch move a sequence across that line on its own. Targeting or offer does that. Where the threshold ladder comes from is a rate definition question and it belongs to our sibling property allboundpros.io, which has not published yet. The second group is anyone who needs cold calling as a core channel. We do not run it and do not design around it, so every cadence here is missing a step that may matter to you.
Plan it instead of arguing about it
The Sequence Cadence Builder takes your prospect count, channels and window and returns a day by day interleaved plan with the coverage window calculated. Free, no signup, no gate on the result.
One division decides which channel can open your sequence. The builder does it and then puts the touches on dates.
If you would rather hand the motion over, the group's agency runs both channels out of one inbox with one shared stop condition, which is what makes a multichannel claim true rather than decorative. If you are comparing named agencies, that belongs on the parent's agency directory, written to a spec that requires naming where the other agency wins. The wider service picture sits at the parent's free GTM diagnostic, which comes back as a plan rather than a pitch.
Frequently asked questions
Is LinkedIn or email better for cold outreach?
Neither, in general. LinkedIn returns several times more per touch, email delivers far more touches, so it depends on your list size and your window. Divide your prospect count by your daily LinkedIn ceiling. If the result exceeds your window, LinkedIn cannot open the sequence.
Which gets more replies, LinkedIn messages or cold email?
LinkedIn, per touch, by a distance. On one client programme in professional services, on the same accounts in the same period, we measured roughly 9% reply rate on LinkedIn DMs against roughly 1.5% on email, with 59% connection request acceptance. Email still produces more total replies, because it places far more touches.
Should I run both channels from day zero?
On small signal triggered lists, yes. On large lists it produces a same day double approach to people who have never heard of you, which reads as a blitz and spends an escalation touch as an opener. Pair the channels deliberately, not by default.
Is cold email dead?
No. The claim usually comes from comparing a ratio measured against replies with a rate measured against sends, which are not comparable. What changed is that authentication and complaint enforcement raised the floor, so the unauthenticated volume approach stopped working.
Do I need Sales Navigator to run LinkedIn outreach?
Not to send, and the seat question is really about search and filtering rather than about sending capacity. Seats, credits and what each tier changes is our sibling property linkedpros.io's cluster, not ours, and it publishes that research properly.
If I can only afford one channel, which do I pick?
Pick the one your capacity can serve for the list you have. Under roughly 500 named accounts, LinkedIn from a real profile. Above a few thousand prospects, email, with LinkedIn as the escalation for whoever shows a signal later.
MultichannelPros is part of the Outbound Pros group, operated by Jānis Plūme.
Last updated: 2026-08-06