Quick answer
Pick channels by where your specific buyer already pays attention, not by what is trendy or what your competitor does. Email is the backbone for most B2B, LinkedIn puts a face to the name, calls cut through for senior buyers and urgent problems. Form a hypothesis from the buyer, test two or three channels small, let positive reply rate set the weighting, then sequence the winners into one story. Cut anything that fails a fair test instead of running it out of habit.
Start with the buyer, not the channel
I am Hlib Storchak. I build and run outbound systems for B2B founders and sales teams, and the single most common reason a campaign underperforms is that the channel was chosen before the buyer was understood. Teams pick email because everyone does email, or LinkedIn because someone on the team likes posting there.
The channel question is really a buyer question. Where does this exact person spend their attention during a working day, and what does contact from a stranger normally look like in that place. A field operations lead and a venture-backed software founder do not live in the same channels, and they should not get the same outreach. That is also why the work starts with an ICP that actually converts: a vague ICP makes the channel question unanswerable.
Four questions that pick the channel for you
Before any test, answer four things about the buyer. First, where do they handle work that arrives from outside the company: an inbox, a phone, a platform feed, or a colleague forwarding something. Second, how many of them exist. A few hundred accounts and thousands of accounts point at completely different channels.
Third, how much is a customer worth, because that sets how many minutes per prospect you can afford. Fourth, how do they normally buy this category: quietly researching, asking peers, or waiting until something breaks. Answer those four and the channel shortlist writes itself, usually down to two candidates and one hypothesis about which leads.
Email is the backbone for most B2B outbound. It scales, it is measurable, it carries detail, and it produces something the recipient can forward to the person who actually owns the budget. If you only run one channel, this is usually it.
Its weaknesses are real. Inboxes in some segments are saturated, and the whole channel depends on infrastructure most teams underrate: authenticated domains, warmed mailboxes, verified lists. Email that never reaches the inbox is not a channel problem, it is a setup problem, and it gets misdiagnosed as bad copy more often than anything else I see.
LinkedIn adds a face to the name and works well for buyers who are genuinely active there: founders, sales and marketing leaders, and anyone whose own job involves being visible. A profile they can check before replying does real work that an email signature cannot.
The constraints are volume and patience. Platform limits keep daily action counts low by design, so LinkedIn is a precision channel, not a volume one, and pushing it hard is how accounts get restricted. It also rewards a warmed-up profile and some visible activity, which takes weeks. The fuller comparison is in cold email vs LinkedIn outreach.
Cold calls
Calls still cut through for senior buyers, urgent problems, and markets where inboxes are saturated. A call also compresses in ninety seconds what an email sequence takes three weeks to establish, which matters when the deal size justifies the minutes.
They demand skill, thick skin, and good phone data, and the economics only work when the account list is small and valuable enough. Calling is not a channel I run, so take this as an honest read on a neighbouring one.
The channels people forget
Three sources beat all of the above on conversion rate and lose badly on volume, which is exactly why they get skipped. Customer referrals and warm introductions convert better than anything cold and almost nobody asks systematically. Communities and industry groups work where the buyer already gathers, provided you turn up as a participant rather than an advertiser.
Event and conference attendee lists give you a real reason to reach out, which is the scarcest ingredient in outbound. So does your own closed-lost and churned list, where the objection is already known and the timing may simply have changed. None of these replace a primary channel. All of them are worth a standing slot in the week.
A channel is right when your buyer already lives there. Not when a competitor uses it, not when a vendor sells it, and not when it is the channel your team happens to enjoy. If you cannot describe, in one sentence, what your buyer is doing in that channel on a normal Tuesday, you do not have a hypothesis yet.
Channel fit by buyer
Use this as a starting hypothesis, not a law. It reflects common patterns, and your own test data overrides it the moment you have any.
| Buyer type | Primary | Support | Why |
|---|---|---|---|
| Software founder | Lives on the platform, checks who you are before replying | ||
| Sales or marketing leader | Processes an inbox daily, forwards what is useful | ||
| Operations or field leader | Cold call | Away from a desk, but reads mail in batches | |
| Senior exec, urgent problem | Cold call | Urgency travels in a voice, inbox is filtered by someone else | |
| Technical buyer, engineering or security | Community | Prefers async and detail, dislikes interruption | |
| Saturated inbox segment | Cold call | Whatever everyone else is doing is the crowded option | |
| Existing network or past buyers | Referral | Warm context beats every cold channel on conversion |
Reading the right mix
Most ICPs respond best to a mix, weighted toward wherever they actually pay attention. The art is in the weighting, not in the list of channels. Email-heavy with a LinkedIn touch suits one ICP. Call-heavy with email follow-up suits another. Running all three at equal weight usually means running all three badly.
A useful discipline: name one primary channel per segment and hold it for the length of a test. Everything else is support. When every channel is primary, nobody owns the result and no number means anything.
Sequencing across channels
Do not run channels in isolation. A connection request, then an email that develops the same idea, then a call that references both, reinforces without repeating. The prospect should experience one person with one point of view, not three tools that happen to share a logo.
The common failure is repetition rather than development. If your LinkedIn message and your email say the same thing in different fonts, the second touch actively costs you credibility. Each touch should add something: a new angle, a relevant detail, a smaller ask. I went deeper on the trade-offs in multichannel vs single-channel outbound.
What each channel costs to run
Channel choice is also a capacity question, and capacity is where most plans quietly break. Here is a model to argue with. Every figure below is an assumption I am showing you, not a benchmark, so swap in your own before you plan anything on it.
Email: assume a warmed mailbox safely handles 30 to 50 cold emails a day, so nine mailboxes across three domains gives roughly 270 to 450 a day. Assume domains and mailboxes at 100 to 300 euros a month, data and verification at 100 to 400, and a sending tool at 50 to 150.
LinkedIn: assume a conservative 15 to 25 connection requests a day per seat, and check the current platform limits yourself rather than trusting any number in an article. That is a few hundred new people a month per seat, plus 30 to 60 minutes a day of human time. Calls: assume 50 to 80 dials a day per caller, with a caller fully loaded at 2,500 to 5,000 euros a month plus 100 to 400 for a dialler and phone data.
The formula for each channel is the same: monthly cost of that channel divided by meetings it produced. One warning on the arithmetic. In a multichannel sequence, attribution is approximate at best, so credit the channel that produced the reply, never count the same meeting twice, and treat the result as a directional comparison rather than a precise figure.
Test before you commit
Run a small test on each plausible channel before scaling any of them. Same offer, same segment, same window, so the only variable is the channel. Let positive reply rate decide the weighting rather than opinion or preference.
Give each test enough volume and enough time to mean something, and decide the window in advance so you cannot move it when week two looks bad. When I take over an account, the first thing I usually find is a channel that was judged on ten days and forty contacts, which is not a test, it is a coin flip with a spreadsheet attached.
When to cut a channel
If a channel produces low positive replies after a fair test, cut it and move the effort into the ones that worked. Running a dead channel out of habit is not neutral: it dilutes attention, inflates your cost per meeting, and makes the reporting harder to read.
Before you cut, though, check the obvious confounders. Was the list equally good in both channels. Did the email actually reach the inbox. Was the LinkedIn profile a stranger's empty page. Plenty of channels get executed badly and then blamed for the outcome.
The mistakes I see most
Choosing channels by preference rather than by buyer. Someone on the team finds calling uncomfortable, so a market that clearly wants a phone gets an email sequence instead, and the result gets blamed on the market.
Adding a channel to fix a message problem. If the offer is not landing by email, it will not land on LinkedIn either. More channels multiply a weak message, they do not repair it. The third mistake is running channels as separate campaigns with separate messages, which turns a coordinated sequence into three strangers contacting the same person in the same week.
How I choose
I start from the buyer, form a hypothesis about their primary channel, test two or three small, and let positive reply rate set the mix. Then I sequence the survivors into one coherent story and hold the weighting until the data says to change it.
Across 2000+ meetings booked for B2B clients, the winning channel was almost always the one the buyer already used, not the one we preferred, and the biggest gains came from weighting an existing channel correctly rather than from adding a new one. When in doubt, go narrower and deeper before you go wider.
Key takeaways
- Pick channels by where your buyer already pays attention, and answer the four buyer questions before running any test.
- Email is the backbone for most B2B, LinkedIn is a precision channel with real volume limits, calls suit senior buyers and urgent problems.
- Referrals, communities, event lists and your own closed-lost list beat cold channels on conversion and lose on volume. Give them a standing slot.
- Name one primary channel per segment. When every channel is primary, no number means anything.
- Sequence channels so each touch develops the idea instead of repeating it in a different font.
- Model capacity and cost per channel from labelled assumptions, decide the test window in advance, and cut what fails a fair test.
FAQ
Should I always use all three channels?
No. Use the ones your buyer actually responds to. More channels is not better if half of them are ignored, and each extra channel costs attention you could have spent making the primary one sharper. Start with one primary and one support, and only add a third when the first two are running properly.
Which channel should a beginner start with?
Email, in most cases. It scales, it is measurable, it fits the majority of B2B buyers, and the feedback loop is fast enough to teach you about your ICP within weeks. The exception is a founder who already has an audience on LinkedIn, in which case start where the attention already is.
Are cold calls dead?
No, they are narrower. They work for senior buyers, urgent problems, small high value account lists, and segments whose inboxes are saturated. They do not work when the addressable list runs to thousands of low value accounts, or when the buyer prefers async contact and treats an unexpected call as an intrusion.
How do I know a channel is not working?
Give it a fair test with enough volume and a window you set in advance, then look at positive reply rate rather than opens or connection accepts. If it stays low while another channel performs on the same list and the same offer, cut it. Before you do, rule out execution problems like deliverability or a thin profile.
How many touches across channels?
A handful of well-spaced, coordinated touches, with each one adding something new. Consistency of message matters far more than raw count. The moment your touches start repeating the same sentence in different places, you are past the useful limit regardless of what the number is.