Most cold outreach is terrible. Not slightly suboptimal — genuinely bad. Generic subject lines, three-paragraph pitches sent to anyone with a vaguely relevant job title, follow-ups that consist of “just bumping this to the top of your inbox.” The kind of email that makes you slightly embarrassed for the sender.
The result is that outbound has a reputation problem it doesn’t entirely deserve. When it’s done properly — with a tight ICP, a genuine value proposition, and sequences that respect the recipient’s time — it generates real pipeline. The problem isn’t the channel. The problem is the execution.
Here’s what actually works.
Why Most Outbound Fails
Before the framework, it’s worth being honest about the failure modes, because most of them are avoidable.
Wrong ICP targeting. Sending to anyone who might conceivably want your product is not a strategy. It’s spray and pray with a CRM. If your sequence goes to 500 people with nothing in common except that they’re vaguely in your industry, your reply rate will be 0.3% and most of those replies will be unsubscribes.
Generic messaging. “I help [role] at [company type] to [vague outcome]” is not a cold email. It’s a template with the merge fields filled in. Recipients can tell the difference between a message that could have been sent to 5,000 people and one that took thirty seconds of personalisation. The former gets deleted.
No sequencing logic. Sending one email and giving up is a waste of your list. Sending seven emails in seven days burns it. Effective outbound sequences have a logic — a reason why each touchpoint exists and why it’s timed the way it is.
Burning the domain. Send too many emails too fast from a new domain with no warmup, and you’ll end up in spam. Once your domain is flagged, you’re done. This is genuinely expensive to recover from.
The Foundation You Need Before You Send Anything
There are three things that need to be in place before you start a sequence. Without them, you’re wasting effort.
A clean ICP. Not a broad one — a specific one. The job title, the company size range, the industry, the specific problem they have that you solve. If you can’t write a one-sentence description of the exact person your best email is going to, you’re not ready. The ICP definition work comes first, every time.
A clear value proposition tied to a specific problem. “We help companies grow revenue” is not a value proposition. “We help SaaS founders between £500k and £2m ARR build a repeatable sales process so they stop being the bottleneck in every deal” is a value proposition. The more specific, the better the response rate.
A CRM that can track what happens. You need to know which emails were opened, which got replies, which contacts converted to calls, and what happened to the deals that came from outbound. Without that, you can’t optimise the sequence and you can’t report on outbound ROI. HubSpot handles this well with proper sequence tracking set up — your outbound activity should feed directly into your deal pipeline.
Building a Proper Prospect List
The list is where most outbound campaigns die quietly. Bad data means you’re emailing the wrong people, hitting dead inboxes, and tanking your deliverability before you’ve even started.
I use Apollo.io for contact enrichment and list building with clients, and it’s genuinely one of the better tools available for B2B outbound. Apollo gives you verified contact data, company filters you can actually use (headcount ranges, technologies used, funding stage), and the ability to build a target list that actually matches your ICP rather than a rough approximation of it. As an Apollo.io partner, I use it in almost every outbound setup I run.
The list-building process should look like this: start with your ICP definition, build company filters in Apollo that match it, identify the right job titles within those companies, verify the contacts, and import them into your sequence tool. That’s it. Do not skip the verification step — undeliverable emails are deliverability poison.
A tight list of 200 highly-qualified prospects will outperform a sloppy list of 2,000 marginal ones every time. Always.
The Anatomy of a Cold Email That Works
Short. Specific. Not a pitch.
The structure that works: a short, specific opening that references something true about their situation (not a compliment — that’s the template move), one sentence on the problem you solve and who you solve it for, one sentence on a relevant proof point or outcome, a low-friction ask.
The entire email should be readable in under 30 seconds. If your cold email requires scrolling, it won’t get read.
The ask matters more than most founders realise. “Would you be open to a 30-minute call to explore whether there’s a fit?” is a big ask from a cold email. “Would it be worth a 15-minute conversation?” is smaller. “Are you the right person to speak to about X, or is there someone else I should reach?” is smaller still — and often generates the most useful replies.
Subject lines: short, plain, not clickbait. “Quick question” outperforms “Transform your revenue function in 2026” by an embarrassing margin.
Multi-Channel: Email Plus LinkedIn
The response rates from email-only sequences have been declining for years. Adding LinkedIn as a second channel — not to pitch, but to create familiarity — meaningfully improves outcomes.
The cadence that works well: connect on LinkedIn (no message, just the connection), send the first email a day or two later. If they accept the connection, that’s a warm signal — follow up on LinkedIn with a short message that matches your email approach. Don’t pitch in the LinkedIn message if you’re already pitching in the email.
The goal of the multi-channel approach is not to bombard — it’s to create two low-pressure touchpoints that together cross the threshold of “this person is relevant and worth replying to.”
Sequences Without Destroying Deliverability
A few hard rules.
Warm up new domains for at least four weeks before sending sequences. Use a warmup tool. Do not send sequences from your main domain if you’re doing significant volume — use a separate sending domain that redirects to your main site.
Limit daily sends per mailbox to 40–50 during warmup, scaling to 80–100 once warmed. Never blast 500 emails on day one.
Keep sequences to four to six steps maximum. Steps one through three are the core sequence — initial email, one follow-up, one breakup email. Steps four through six can include a LinkedIn touchpoint or a different angle on the value proposition. More than six steps and you’re pestering people who’ve made a decision.
What Good Response Rates Actually Look Like
Founders consistently overestimate what outbound delivers. Here are realistic benchmarks.
For a well-targeted sequence with good copy: open rates of 40–60%, reply rates of 3–8%, positive reply rates (actual interest, not just “remove me”) of 1–3%. For every 100 people in your sequence, expect one to three useful conversations.
That sounds low. But if your average deal is £30,000 ARR, and you’re generating four qualified conversations per month from outbound, that’s meaningful pipeline. The maths only works if you’re running the sequences consistently and if your qualification is tight enough to convert those conversations efficiently.
SDR vs Doing It Yourself
If you’re at early stage, you probably can’t justify an SDR yet. The typical fully-loaded cost of a junior SDR in the UK is £45,000–£55,000 per year, plus the time to manage, coach, and ramp them. That investment only makes sense once outbound is proven as a channel — i.e., you’ve run it yourself, you know what works, and you’re generating consistent pipeline that justifies dedicated resource.
Do it yourself first. Build the playbook. Prove the channel. Then hire the person to run the system you’ve built rather than asking them to build it from scratch.
The foundation of any effective outbound programme is a genuinely specific ICP — without it, everything else is guesswork. If you haven’t nailed that yet, defining your ICP properly is the right place to start. And if you want to understand how AI tooling is changing what’s possible in terms of prospect research, personalisation at scale, and sequence optimisation, the AI RevOps page covers what’s actually useful versus what’s just noise.