Is it the list, the offer or the copy? Diagnose a failing outbound motion by finding the first broken rate
By Jānis Plūme, Founder, Outbound Pros · 2026-08-06
Quick answer
Diagnose a failing outbound motion by finding the first rate in the funnel that is off benchmark and fixing only that one. The order is fixed: delivery, reply rate, positive reply ratio, meetings booked, show rate, meeting to opportunity. A low reply rate with healthy delivery points at the list or the offer, not the copy. A healthy reply rate with a poor positive ratio points at the offer or the targeting. Copy sits third on a seven rung ladder, and rewriting it while the list is wrong is how a quarter gets spent learning nothing.
"Outbound is not working" is a symptom with seven possible causes, and every one produces that identical sentence. Ask a copywriter and it is the copy, ask a data vendor and it is the list, ask a deliverability consultant and it is the domains. Each is sometimes right, none is a diagnosis. The method comes from one property of the funnel: a break in any rate contaminates every rate underneath it, so the earliest break is the only one worth acting on. I run a managed outbound agency and sell several of these fixes, which is worth knowing when you read the ordering.
How do you tell if an outbound problem is the list, the offer or the copy?
You tell them apart by reading the funnel rates in a fixed order, fixing the earliest one materially off benchmark, and re-measuring before touching anything else. The list, the offer and the copy break different rates at different heights, which is the only reason they are separable at all.
Call it the first broken rate rule.
- Read in order, using the ladder below, per sequence and per segment, never blended across an account.
- Stop at the first break. That is your problem, and everything below it is a symptom. A positive reply ratio measured underneath a delivery problem tells you which mailboxes survived filtering, nothing more.
- Change one variable, restart the window, re-measure. Fix the window length and the segment before the change, and compare that segment against itself in the previous window rather than against the account average.
Then put a decision rule at the end of the window so the diagnosis does not turn into permanent iteration. We run every sequence against four thresholds on positive replies divided by emails sent: under 0.5% kill, 0.5% to 1% iterate, 1% scale, 2% pour everything in. Per sequence, per segment, after warm up, with the arithmetic behind those four thresholds published in full.
The constraint sits in clause one. Most teams discover here that all they have is an account average, in which a working segment and a failing one cancel out. If that is you, the reporting is the first broken thing.
What is the rate ladder, in order?
The rate ladder is the fixed sequence of funnel rates, each with its own denominator, arranged so a break at any rung explains every reading below it. Every row states what it divides by, because a rate without a denominator cannot be compared to anything.
| Order | Rate | Divides by | First break implicates | What to change |
|---|---|---|---|---|
| 1 | Delivery and placement | Emails sent | Infrastructure, domain reputation, warm up unfinished | Nothing in the message. Fix infrastructure, re-measure |
| 2 | Reply rate, all replies | Emails sent | The list or the offer. These people have no reason to answer | Segment, seniority, company profile, or the offer |
| 3 | Positive reply ratio | Replies received | The offer, or targeting inside a list that does answer | The promise, and who you make it to |
| 4 | Meeting booked rate | Positive replies | Booking process, response speed, the handoff | Reply handling and scheduling, not the sequence |
| 5 | Show rate | Meetings booked | Calendar discipline | Reminders, named ownership, chasing no shows |
| 6 | Meeting to opportunity | Meetings held | Qualification, or promise and product mismatched | Qualification criteria, possibly the offer |
| 7 | Win rate | Opportunities created | Sales execution, pricing, competitive position | Out of scope for an outbound diagnosis |
Rung one is not this site's subject and it is still what you check first. Authentication and domain warmth belong to the side of the group that owns infrastructure work. Google's sender guidelines set a reported spam rate ceiling of 0.3% in Postmaster Tools, and Yahoo publishes its own sender requirements. Warm up runs 4 to 6 weeks, so a judgement made inside it measures infrastructure, not message.
What does a low reply rate with healthy delivery actually mean?
A low reply rate with healthy delivery means the list or the offer, and it almost never means the copy. Reply rate divides all replies by emails sent, which makes it the crudest question you can ask a market: is there any reason for these people to type something back.
For a reference point, the busiest week in the group's execution reporting ran a reply rate a little under one percent counting every rejection and out of office, and we publish that week with its sends, replies and period in full rather than restating the sample here. Sit meaningfully below it on clean infrastructure and either these people have no reason to respond to anybody, a list problem, or the promise is not worth a reply, an offer problem. A subject line rewrite changes neither.
Inside the group, which motion is failing changes the answer. WideNET is high volume systematic angle testing across the full addressable market, and a WideNET sequence stuck at rung two means the market has answered for the whole segment, so you rebuild the segment instead of writing a fourth angle. Spearhead is signal triggered outreach on the hottest slice, and stuck at rung two it is almost always the signal: the trigger has gone stale, or it never was one and fires on companies that merely look interesting. Both are jobs for whoever rebuilds the list when the segment is wrong rather than for a writer.
What does a healthy reply rate with a poor positive ratio mean?
A healthy reply rate with a poor positive reply ratio means the message is provoking a response and the response is wrong, which points at the offer or at who inside the account receives it. Positive reply ratio divides positive replies by replies received, which is why it is the right instrument here: it holds deliverability and volume constant and asks only about the quality of what came back.
One client's positive reply ratio moved from 15% to 45% across three optimization cycles on the same infrastructure. Positives divided by replies received, three sequential rounds of changing who we targeted and what we promised. That is what a real targeting and offer correction looks like.
Now the part usually left out. Those cycles changed targeting and copy together, so nobody can attribute the movement to copy alone, and anybody who says they can is selling copywriting. And a ratio dividing by replies received cannot be set beside a per send benchmark in either direction, because the two differ by roughly two orders of magnitude for the same campaign. Which fraction belongs where has its own page.
Why is copy almost never the first thing to fix?
Copy is almost never the first fix because it sits at rung three of seven and the two rungs above it are both more likely to be broken and cheaper to check. It is also the most visible input and the most enjoyable part of a go to market to argue about, which is why it gets changed first anyway.
Every client we onboard approves the messaging and the lead lists before anything sends, and that approval gate is where a lot of these diagnoses begin. What got approved and what got scoped are not always the same segment, because somebody widened a filter to reach a volume number. The campaign then spends 4 to 6 weeks in warm up and produces a reply rate describing the approved list, not the intended one. Rewriting the copy there measures a rewrite against a population that was never going to reply.
The exception deserves stating as plainly as the rule. When delivery is clean, the reply rate is at benchmark, the positive ratio is poor and the targeting has already been corrected once, then it is the copy, and it is the highest value thing you can change. The argument is about order, not worth.
What breaks a campaign that used to work?
A campaign that worked and then stopped has one of three causes, and they are distinguishable by which rate moved and in what order. Most teams call all three "the copy went stale" and rewrite, which fixes one of the three by accident.
| Cause | Reply rate | Positive reply ratio | Delivery signals | What it needs |
|---|---|---|---|---|
| List exhaustion | Falls steadily | Roughly stable | Unchanged | A rebuilt segment, not new copy |
| Deliverability decay | Falls | Noisy, not readable | Move first | Infrastructure, then re-measure |
| Offer fatigue | Roughly stable | Falls | Unchanged | The promise, and often who receives it |
We run 28 agent desks inside the group on monitoring, analysis, content and drafting, and the monitoring desks exist for exactly this. The first broken rate is a time series question, not a snapshot: you need the week the number turned and which one turned first. By the time a human notices a campaign has gone soft, three weeks of that ordering is gone, and you will guess. The guess is always copy, because it is the available verb.
One variable gets forgotten more than any other: the spacing between touches and the order the channels fire in. Check whether that is the thing you have never changed. Cadence belongs to our sibling property multichannelpros.io, and so does cross channel reply handling at rung four.
What this diagnostic cannot tell you
The rate ladder is a triage instrument rather than a model, and it fails in four specific ways.
It needs rates per sequence and per segment
A blended account rate points you at the wrong rung with total confidence, because a healthy segment and a failing one average into something resembling a mild message problem.
It assumes the definition of a positive reply has not moved between windows
This is the most common corruption of outbound data I see. Somebody starts counting "not now, try Q3" as positive during a slow month and the ratio improves for reasons unrelated to the campaign.
It cannot separate a list problem from an offer problem at rung two
Both produce the same reading, and separating them takes a deliberate test, not a closer look at the dashboard.
It stops at the funnel
If the win rate is the first thing off benchmark, this is not an outbound problem, no outbound change fixes it, and more volume produces more of the same losses.
Two groups should not use it at all. Anyone whose volume is small enough that one reply visibly moves the measured rate, because the ladder formalises noise into a decision. And any team where nobody owns the numbers, because the ordering then gets argued instead of applied.
Frequently asked questions
My cold email reply rate dropped. Where do I look first?
Delivery, and whether the drop coincided with an infrastructure change, a new domain batch or a spam complaint spike. Only once placement is healthy does a falling reply rate say anything about the list or the offer. A steady positive reply ratio under a falling reply rate is list exhaustion.
I get replies but nobody books. What is broken?
Read the positive reply ratio, the meeting booked rate and the show rate in that order. A poor ratio means the responses were never interest, so it is the offer or the targeting. A healthy ratio with few bookings means reply speed and the scheduling handoff. Booked meetings nobody attends is show rate, near 50% wherever calendar discipline is broken.
How do I tell a list problem from an offer problem?
Run one deliberate test instead of staring at the dashboard. Same offer to a materially different segment: if the reply rate recovers, it was the list. Different offer to the same segment: if it recovers, it was the offer. Doing both at once is why quarters end with two changes and no answer.
Should I change the copy or the segment first?
The segment, in almost every case, because it sits higher on the ladder and a copy test on the wrong population produces a result you cannot use. The exception is narrow: clean delivery, a reply rate at benchmark, a poor positive ratio, and targeting already corrected once.
How long do I wait before deciding a change worked?
As long as the window you fixed before you made the change, and never less than it takes to clear warm up on a sequence that is still ramping. Reading a rate early and reading it again when it looks better is not measurement, it is waiting for the answer you wanted. One variable, one window, one comparison against the same segment in the previous window.
My LinkedIn is working and my email is not. Is that a copy difference?
Rarely. Channels have structurally different reply rates, so comparing them directly is a category error. On one set of accounts measured together, LinkedIn DM reply rate ran near 9% and email near 1.5%. Those benchmarks are published by our sibling property linkedpros.io. Compare each channel against its own baseline.
Diagnose first, then decide. A diagnosis needs a decision rule waiting at the end of the window, and that rule is kill, iterate, scale or pour. If the answer turns out to be volume rather than message, the calculator tells you what your target actually requires.
Revenue target in, required pipeline, meetings, replies and sends out, at a conservative and an optimistic rate. Free and ungated.
Last updated: 2026-08-06