Prospecting channels: what actually works in 2026

A channel is a route to starting a conversation. The core truth of modern prospecting: no channel works alone, and average performance everywhere has declined as buyers drown in AI-generated outreach. What still works is relevance — provable knowledge of the prospect's actual situation — delivered across channels in sequence.

Realistic benchmarks (orders of magnitude — quote as ranges in interviews):

• Cold email (unsolicited email to a stranger): reply rates of 1–5% are normal; well-researched, personalised sends reach 5–10%. A good cold emailer books perhaps 1 meeting per 100–200 sends; a great one does better on smaller, better-chosen lists.
• Cold calling (phoning someone who isn't expecting you): roughly 5–10 dials per live conversation; of conversations, perhaps 5–15% agree to a meeting — so ~100 dials → 1–2 meetings. Brutal per attempt, but it's the only channel with a live human reaction, and in saturated UK healthcare inboxes a good call stands out because so few people make them well.
• LinkedIn / social selling — using LinkedIn to build visibility and start conversations: sequences convert about like email, but the compounding value is warmth — engaging usefully with a prospect's posts before you pitch. Buyers check who you are before replying to anything, so your profile is part of every channel.
• Referrals — introductions through people who know you: conversion to meeting is dramatically higher than any cold channel (30–70%+), because trust transfers. This is why experienced sellers invest in relationships and ask for introductions long before any sequencing tool comes out.
• Events and conferences — expensive per contact but high quality; a 10-minute in-person conversation can replace a 12-touch cadence. If you've ever staffed a stand or represented an organisation at an industry event, you already know the motion.
• Inbound — prospects who come to you (downloaded a whitepaper, requested a demo): converts best of all, but marketing controls the volume. AE interviews focus on outbound (you initiating) because that's the skill being hired.

Voice from the field

"The brutal fact is the number one reason for failure in sales is an empty pipe and the root cause of an empty pipeline is the failure to consistently prospect."
— Jeb Blount, Fanatical Prospecting (2015)

Multi-touch sequences (cadences)

A sequence or cadence is a planned series of touches to one prospect across channels over time — because a single touch of anything is nearly invisible. A typical 2026 B2B cadence commonly runs 8–16 touches over 2–4 weeks — a practitioner rule of thumb, not a measured statistic:

• Day 1: personalised email
• Day 2: LinkedIn connection with a short note
• Day 4: call (voicemail if no answer)
• Day 6: one-line email bump
• Day 9: comment on their LinkedIn post
• Day 11: call
• Day 14: email with a new angle
• Day 18: call
• Day 21: "break-up" email closing the loop (these often get disproportionate replies) → nurture list.

Two principles:

• Every touch adds something new (an insight, asset or angle — never "just checking in").
• The cadence is a system, not a mood — run consistently, measured and iterated, like instrumenting a funnel in software.

How to write a cold email that gets replies

Structure — short enough to read on a phone (60–120 words):

1. Subject line: short and specific — under ~40 characters, roughly 2–7 words; casual lower-case often outperforms formal ("reporting backlog at [trust acronym]" beats "Revolutionise Your Radiology Workflow"). No clickbait.
2. Personalised opener (1–2 lines): proof you know them — board papers, published strategy, a talk, a job advert. This line earns the rest being read; generic flattery reads as template.
3. Problem/insight (1–2 lines): a problem their peer group demonstrably has, in their language — not your features.
4. Credibility (1 line): one proof point — a named peer customer, a number, a published result.
5. Call to action (CTA) — the specific thing you ask them to do. Low-friction beats high-friction: "worth 15 minutes?" demands commitment; "is this on your radar this year?" invites a one-line reply. Interest-based CTAs (about the problem) outperform meeting-based CTAs (asking for time) on first touch.

Personalisation comes in tiers:

• account-level (the organisation)
• person-level (this individual)
• timing-level (this month — a compelling event)

Aim for two of three. Everyone now has AI that generates fake-personalised email, and buyers can smell it; what can't be faked is a genuinely correct observation about their situation.

Field note

Treat any benchmark a sequencing-tool vendor publishes with suspicion — their business depends on outbound looking healthy. Open rates in particular have been near-meaningless since Apple's Mail Privacy Protection began pre-fetching emails, so serious teams measure replies and booked meetings only. And be ready for the unglamorous truth: even at a good 5% reply rate, most of those replies are "no" or "not now" — a well-run programme is defined by how gracefully it handles the rejections and the silence, not by its best week.

Worked example 1 — NHS trust CIO, clinical AI

Subject: reporting backlogs & your 26/27 digital plan

Hi [name],

Your trust's April board papers flag radiology reporting turnaround as a risk to the 28-day Faster Diagnosis Standard — and the digital strategy commits to "evaluated AI adoption" this year.

That combination is where we work: [Product] triages chest imaging so urgent findings are reported first. [Comparable trust] cut median time-to-report for critical findings by X% in a published evaluation, on existing PACS infrastructure.

Is AI triage part of the plan for this financial year, or later? Happy to share the evaluation either way.

[signature]

Why it works:

• The opener proves real research (board papers are public — reading them is a genuine differentiator).
• The problem is theirs, a standard they're measured on.
• One proof point with a number.
• The CTA asks about timing, not for a meeting.

Worked example 2 — biotech lab operations lead, sample management

Subject: freezer inventory ahead of the [site] expansion

Hi [name],

Saw you're hiring two lab techs for the new [site] facility — congrats. Scaling a biobank across sites is usually the point where spreadsheet-based sample tracking starts costing real money: mislocated samples, failed audits, freezer space bought that isn't needed.

We help teams like [peer biotech] manage sample inventory across sites — they recovered ~30% of freezer capacity in the first quarter just from visibility.

How are you handling sample tracking across the two sites today?

[signature]

Why it works:

• The trigger (job adverts = expansion) is timing-level personalisation.
• Three concrete pains rather than "inefficiency".
• The CTA is a discovery question — any answer starts a conversation.

3. Cold calling: a structure you can say out loud

Cold calls fail when the caller sounds like a script and the prospect feels ambushed. The fix is a structure that acknowledges the interruption and earns each next 20 seconds.

1. The opener

Name and company plainly — no "how are you today?" theatre. The strongest recent pattern is the permission-based opener: acknowledge the cold call and ask for a moment.

"Hi [name] — [your name] calling from [company]. I'll be honest, this is a cold call — can I take 30 seconds to tell you why I rang, and you can tell me if it's relevant?"

It trades a little momentum for a lot of trust: the prospect keeps control, so they don't need to seize it by hanging up. Most people, asked politely for 30 seconds, grant them.

2. The 30-second problem statement

Spend them on their problem, not your product:

"The reason for the call — we work with lab ops leads at mid-size biotechs, and the pattern we keep hearing is that sample tracking lives in spreadsheets until the day a sample can't be found for a regulatory audit. I saw you're scaling to a second site, which is usually when that bites. Is that close to your world, or have you got it solved?"

Anatomy:

1. who we work with (their peer group)
2. the problem in their words
3. why I called you specifically (the trigger)
4. an easy question

The escape hatch ("or have you got it solved?") gives permission to say no — which paradoxically increases honest yeses.

3. The goal is a conversation, then a next step

Five minutes of genuine discovery is a success even if the meeting gets booked by email afterwards. Ask two or three real questions; don't pitch features.

4. Handling "just send me an email"

Usually a polite brush-off, not a request. Never fight it; use it:

"Happy to. So I don't send you generic rubbish — is the two-site sample tracking thing actually a live issue, or is something else at the top of your list?"

You agree (lowering resistance), then ask one qualifying question. Either you learn what matters (and your email becomes relevant), or you disqualify in sixty seconds. A useful follow-up — "who else would need to see it?" — quietly maps the buying committee.