The 3% positive-reply line: when to fix a campaign

We run cold email campaigns against a simple rule: expect roughly 4% positive replies, and treat anything below 3% as a fault condition — stop scaling and start diagnosing. The 3% line matters because it converts a vague feeling ("outbound isn't really working") into a decision with a defined next action.

What counts as a positive reply?

A reply that moves the conversation forward: expressed interest, a question about what you do, a referral to the right colleague, or an agreement to talk. "Not right now, try me in Q4" counts — it is a live prospect with a date attached. Out-of-office, unsubscribes, and flat refusals do not.

Measured per prospect contacted, not per email sent. The full measurement logic — and why total reply rate is a decoy — is in what's a good cold email reply rate in UK B2B, and the machine that produces the number is described end-to-end in the UK B2B outbound playbook.

Why draw the line at 3%?

Because a threshold forces behaviour. Without one, a mediocre campaign runs for months on hope: the numbers are "a bit soft", someone means to look at the copy, nothing changes. With one, the rule is mechanical — when the campaign has contacted a meaningful sample and sits below 3% positive, it goes into diagnosis; when it sits at 4% or above, it earns more volume.

One caution: do not judge a campaign on 40 sends. Below a couple of hundred prospects contacted, the rate is mostly noise. Let the sequence of four emails over 14 days complete for a real cohort before ruling.

What do you check first when a campaign is below the line?

Diagnose in a fixed order, cheapest and most fundamental first. Copy comes last, because it is the most enjoyable thing to fiddle with and the least likely culprit.

1. When bounces are high or replies are near zero, check deliverability first. A domain that lands in spam makes every other question irrelevant. Check warm-up scores, authentication, and daily volume — each inbox should be sending 25–40 emails a day, not hundreds.
2. When deliverability is clean but replies are flat, check the list. Is the ICP actually precise, or is it "UK companies, 5–50 staff, various"? Are contacts verified and current? A decayed or broad list caps the rate no matter what the emails say.
3. When the list is sound, check the offer. Are you naming a problem this sub-vertical recognises, in their language, with a specific outcome? Generic value propositions produce polite silence.
4. Only then touch the copy. Length, first line, the ask. Change one variable, run a fresh cohort, re-measure.

Which tool tells you which of these is failing — and what each layer of the stack costs — is laid out in the outbound stack: what each piece does and costs.

When do you fix, and when do you kill?

Fix when one link in the chain is identifiably broken: a spam-folder problem, a stale list segment, an offer that missed the sub-vertical's actual pain. Those are repairs with a known procedure.

Consider killing the campaign — the campaign, not the channel — when two full diagnostic passes have not moved the rate. Usually that means the targeting premise was wrong: this sub-vertical does not feel this problem strongly enough to reply. The correct move is a new campaign against a different segment or a different problem, not a third rewrite of the same email. Campaigns are cheap to start; months of denial are not.

What should you do above 4%?

Scale — but by addition, not by pushing. Keep each inbox at 25–40 emails a day and add mailboxes and parallel campaigns to grow volume; a working campaign is an asset to protect, not a dial to crank until deliverability breaks.

Why have a numeric line at all?

Because without one, every campaign judgement routes through the founder's gut, and the founder becomes the bottleneck for a decision a spreadsheet could make. Thresholds are how you delegate judgement to a system — the broader argument is in the Founder-as-Bottleneck Report. A campaign with a defined fault condition can be run by someone other than you. A campaign judged by feel cannot.

---

Next step: the Growth System Audit — £450, seven days, credited against any build — maps where your growth system leaks and what to build first.