A typical outreach dashboard shows emails sent, opens, clicks and total replies. All four move every week. None of them says, on its own, whether you are starting conversations with people who can buy.
The outreach metrics that matter form a chain: people contacted, messages delivered, replies, positive replies, meetings booked, meetings held, and the opportunities that follow. Each link has a definition and a formula, and when one weakens it points to a different fix from the others.
You don’t need a BI tool to read that chain. A spreadsheet, definitions that never change and one hour a week will do.
The outreach metrics chain#
Treat outreach as a series of conversions. A person enters the chain when you first contact them, and each later stage is a subset of the one before it. Measure every stage against the stage above and a weak number tells you where to look.
| Stage | What counts | What a weak number usually points to |
|---|---|---|
| Contacted | Unique people who received at least one step of a sequence | Not enough verified, in‑profile contacts |
| Delivered | People whose messages did not bounce | List quality or sender setup |
| Replied | People who sent any human reply | Targeting, relevance or the offer |
| Positive reply | Replies that show interest or ask a buying question | Message fit or the wrong buyer title |
| Meeting booked | Positive replies that turned into a calendar event | Response speed and scheduling friction |
| Meeting held | Booked meetings that took place | Confirmation and reminders |
| Opportunity | Held meetings that became a qualified deal | Qualification before the meeting |
Two counting rules keep the chain honest. Count people, not messages: someone who gets a four-step sequence and replies once is one contact and one reply. And count human replies only. Out-of-office notices, auto-replies and bounces are events to process, and they stay out of the reply column.
The later stages lag behind. Replies come within days of a send, meetings within a week or two, opportunities only after the first call. Read the early stages weekly and the later ones monthly, always for a fixed cohort of people contacted in the same period. Mix cohorts and a slow month for meetings will look like a bad month for messaging.
How to calculate reply rate and other cold email metrics#
Write these formulas down once and leave them alone. A change in definition looks exactly like a change in performance. If you ever have to change one, log the date next to it.
Delivery rate = delivered people ÷ contacted people × 100
Bounce rate = bounced addresses ÷ addresses sent to × 100
Reply rate = people with a human reply ÷ delivered people × 100
Positive reply rate = positive replies ÷ delivered people × 100
Reply-to-meeting rate = meetings booked ÷ positive replies × 100
Show rate = meetings held ÷ meetings booked × 100
Add one summary figure that a CEO can read without a briefing: meetings held per 100 people contacted. It folds every stage into a single number and lets you compare campaigns of very different sizes.
A worked example, with illustrative numbers. Say you contact 400 people and 388 are delivered. Twenty-four reply, ten of them positively. Six of those ten book a meeting and five turn up. That gives a reply rate of about 6.2%, a positive reply rate of about 2.6%, a reply-to-meeting rate of 60% and a show rate of about 83%. In summary form, 1.25 meetings held per 100 people contacted.
A single reply rate would have hidden the weak spot. Four of the ten interested people never booked, which makes this a process problem more than a messaging one. The guide to reply-to-meeting conversion works through that stage step by step.
Why open rates are unreliable for cold email#
Open rate used to be the first number anyone checked. Today it is the least reliable one on the dashboard. An open is logged when a tracking image loads, and an image loading is a different event from a person reading.
Apple’s Mail Privacy Protection shows the problem most clearly. Apple’s own description of the feature says it downloads remote content in the background by default, regardless of whether the recipient engages with the email. For anyone using it, a tracked open can fire with nobody reading anything. Image blocking pushes the number the other way, and security scanners add noise of their own.
For decisions about outreach, treat four numbers as vanity metrics. Opens are inflated by automatic downloads and deflated by image blocking. Emails sent measure effort, and sending more to the wrong people drags down every rate beneath it. Total replies flatter you until you strip out the out-of-office notices, the unsubscribes and the “not interested”. Link clicks can come from security tools that scan links before delivery, with no person involved at all.
Use replies as the attention signal instead. A reply means someone read the message and decided to answer it.
Guardrail metrics: bounces, complaints and unsubscribes#
Some metrics don’t measure success. They measure whether you get to keep sending. Trouble here shows up later as a falling reply rate on every campaign you run.
Bounce rate is the list-quality signal. When it climbs, contacts usually weren’t verified before sending, or the data source has gone stale. Track it per list source as well as overall, since one bad export can hide inside a healthy average.
Spam complaint rate has an explicit ceiling. Google’s email sender guidelines ask senders to keep the spam rate reported in Postmaster Tools below 0.10% and to avoid ever reaching 0.30% or higher. Outreach that draws complaints loses the inbox long before it loses the prospect.
Unsubscribe rate measures relevance. A spike right after a message change means the new version reached people who shouldn’t have been on the list, or told them something they didn’t want to hear. Honor every unsubscribe at once, whatever it does to the numbers.
Review guardrails per inbox and per sending domain, and pause anything that crosses your own thresholds. The guide to email deliverability for cold outreach covers the setup that keeps all three low.
Read every metric by segment#
An overall reply rate averages very different groups. One segment can be producing every meeting while another produces the complaints, and the campaign total looks steady the whole time. Split each metric by the things you control:
| Segment by | Question it answers |
|---|---|
| Target segment (industry, size, region) | Which part of your ideal customer profile responds |
| Buyer title and seniority | Are you writing to the person who owns the problem |
| Message version | Which angle or offer earns positive replies |
| Sequence step | Do follow‑ups add replies or only volume |
| Channel | Which mix of email, LinkedIn and WhatsApp works for this audience |
| Sending inbox and domain | Is one sender dragging delivery or complaints down |
| List source | Which data source has the cleanest, most relevant contacts |
Take the 400 people from the example above. Say half are operations leaders at manufacturers and half are finance leaders at the same companies. If eight of the ten positive replies come from operations, the finding isn’t a 2.6% positive reply rate. Operations owns the problem, finance doesn’t, and next month’s list should be built around that.
Small segments mislead. A gap of one or two replies between groups of fifty is usually noise. The article on building your own cold outreach baseline shows how to tell a real difference from a lucky week.
Segmenting also exposes trade-offs. A shorter message can lift the reply rate while positive replies fall. A broader list can add two meetings and double the complaints. Neither shows up until the whole chain is read per segment.
A weekly outreach metrics review#
Numbers help only when someone looks at them on a schedule and then acts. Thirty to sixty minutes a week is enough for most teams. Keep the order fixed:
- Guardrails first: bounces, complaints and unsubscribes per inbox and domain. Pause anything unhealthy before the conversation moves on.
- Read the chain for last week’s cohort (contacted, delivered, replied, positive, booked, held) and find the weakest link against your own baseline.
- Segment that link. Is the drop in one industry, one title, one message version or one sender?
- Read ten replies in full. The numbers show where the problem is and the replies show why. Look for objections that repeat and questions you could answer better.
- Pick one change and write it down with the date. With one variable per test, next week’s numbers mean something.
- Check the handoff. Every booked meeting should be in the CRM with its email thread attached. A meeting nobody can find is a lost meeting.
After a quarter of these reviews you have a dated log of what you changed and what happened next. Your next campaign starts from that log.
Rhenyx watches bounce and complaint rates on every mailbox and slows sending before your domain takes the hit, and puts every booked meeting in your CRM with the whole conversation attached. The chain from contacted to meeting is recorded in one place. Rhenyx is the outreach platform for B2B teams: you approve the plan, and it runs outbound from your website to a booked meeting. See How the Platform Works