You sent the campaign, the dashboard is full of numbers, and you still can't tell whether it worked or what to fix. That is the real problem with campaign metrics: there are plenty of them, but only a few tell you what to do next.
This guide gives you a short list of numbers worth watching, explains which ones lie, and walks you through a routine you can run on your current campaign this afternoon.
Start with the one number that pays you
Before looking at anything else, decide what a win is. For most cold outreach it is a positive reply: someone saying yes, asking a question, or agreeing to a call. Not opens, not clicks.
Write this down for each campaign:
- Goal: for example, booked calls with local dental clinics
- Win metric: positive replies per 100 emails delivered
- Secondary metric: meetings booked
Every other number is useful only because it explains why that win metric is high or low.
The metrics worth tracking, in order
Read your numbers top to bottom, like a funnel. A problem higher up makes everything below it meaningless.
| Metric | What it tells you | What to fix if it's bad |
|---|---|---|
| Bounce rate | Is your list clean? | Verify and clean the list |
| Unsubscribes | Is the targeting or tone off? | Narrow the audience, soften the ask |
| Reply rate | Does the message land at all? | Offer, first line, targeting |
| Positive reply rate | Are the right people interested? | Offer and audience fit |
| Meetings booked | Does interest turn into action? | Your call to action and follow-up speed |
Bounce rate
This is the first thing to check because it affects your sender reputation. A common rule of thumb is to keep bounces very low, a few percent at most. If yours is higher, stop sending and fix the list before anything else. Our lead list cleaning checklist covers the steps.
Reply rate and positive reply rate
Track both. A high reply rate full of "not interested" and "remove me" means your message is clear but the offer or audience is wrong. A low reply rate with mostly positive replies means you are close and should test subject lines or the first line.
Tag every reply by hand: positive, neutral, negative, out of office. It takes minutes and gives you the most honest data you have.
Meetings booked
If positive replies are coming in but meetings are not, the issue is usually after the reply: slow responses, vague next steps, or too many questions. Answer interested people quickly with one clear time suggestion.
The metrics to treat with caution
Open rate
Opens are tracked with a tiny image. Some mail apps load images automatically, which records opens that never happened. Others block images, which hides real opens. So open rates can be inflated or deflated, and you can't tell which. Read more in how accurate open tracking really is.
Use opens only to compare two versions of the same campaign sent to similar lists at the same time. Never use them as proof a campaign works.
Click rate
Clicks are more reliable than opens, but security scanners at some companies click links to check them. A burst of clicks seconds after sending is often a bot. Also, putting links in a first cold email can hurt deliverability, so many senders leave them out entirely, making click rate irrelevant.
A 30-minute campaign review you can run today
- Pull the numbers for one campaign only. Mixing campaigns hides problems.
- Check bounces first. If they are high, pause sending and clean the list. Stop here.
- Count replies and tag them. Work out positive replies per 100 delivered.
- Read every negative reply. Look for patterns: wrong person, wrong timing, already have a provider. Each pattern points to a fix.
- Check reply timing. If replies cluster at odd hours, your send times may be off for the recipient's location. See time zones in cold email.
- Look at each step separately. Which step in your sequence produced replies? If step three does nothing, rewrite or cut it.
- Pick one change. Only one. Note what you changed and the date.
Change one thing, then wait for enough data
The fastest way to learn nothing is to change the subject line, offer and audience in the same week. Change one variable per test.
As a rule of thumb, wait until each version has gone to at least a hundred or so delivered contacts before judging. Small samples swing wildly; two replies versus four can look like a doubling and mean nothing.
Good things to test, roughly in order of impact:
- Audience: a tighter niche or a different city
- Offer: what you are asking for and what they get
- First line: something specific to the business, such as its category or city
- Subject line: short and plain
Audience often matters more than wording. Personalising with the owner's actual name helps too; here is how to find business owner names.
Use behaviour to shape follow-ups, carefully
Open and click signals are imperfect, but they are still useful for routing follow-ups rather than for judging success. For example, someone who clicked a link might get a more direct follow-up, while someone who showed no activity gets a different angle or a fresh subject line.
In GetClient, the flow builder lets you branch on If Opened, If NOT Opened, If Clicked and If NOT Clicked, with delays in days, hours or minutes. Just remember the branch is yes or no, and because open data is noisy, treat it as a nudge, not a verdict. Replies land in one Inbox, which makes the manual tagging step above quicker.
You can also combine engagement with fit to decide who to call first. The ideas behind lead scoring apply here.
Keep a simple log
A spreadsheet beats memory. One row per test:
- Campaign and date
- What changed
- Delivered, bounces, replies, positive replies, meetings
- What you learned
After a few weeks you will see which audiences and offers actually work for you, which is worth more than any benchmark from someone else's campaigns.