What makes an email campaign 'good' if not its open rate or click rate?

A good campaign is one with sound logic from end to end: a clear reason for existing, an audience defined by behaviour rather than just a list segment, a specific action it expects the reader to take, a success criterion that was set before it sent, and a process for feeding what is learned back into the next campaign. Benchmarks can be hit by accident; logic cannot.

What is the difference between a campaign goal and campaign intent?

Intent is the underlying reason the campaign exists — what business or relationship problem it is solving. A goal is usually a number attached after the fact, like 'get a 25% open rate.' Campaigns built around intent tend to hold together structurally; campaigns built around a goal number often have no clear reason to exist beyond hitting that number.

How specific should audience targeting be for an email campaign?

Specific enough that you could describe the recipient's likely situation in one sentence — for example, 'browsed twice but did not purchase in the last 14 days' rather than 'everyone on the list.' Behavioural and lifecycle-stage targeting consistently outperforms broad demographic or list-wide targeting because the message matches what the recipient is actually doing.

What is 'expected behaviour' in an email campaign?

Expected behaviour is the single action the campaign is designed to produce — a click, a reply, a purchase, a re-engagement. Every design and copy decision in the email should serve that one expected behaviour. Campaigns that try to produce several behaviours at once (click here, also reply, also follow us) usually produce none of them well.

Why do some teams skip the post-send learning loop?

Usually because the team moves straight to planning the next campaign once a send is out, and post-send review gets treated as optional. Without it, every campaign is planned from instinct rather than from what the last one actually taught the team, and the same mistakes get repeated indefinitely.

What a 'Good Email Campaign' Actually Looks Like in 2026 (Real Breakdown)

A "Good Campaign" Is Not a Number

Ask a team what a good email campaign looks like and they will quote a benchmark. Twenty-five percent open rate. Three percent click rate. Two percent conversion. Numbers borrowed from an industry report that has no idea what their audience, offer, or list actually looks like.

Benchmarks are comparison tools. They are not a definition of quality, and they can be hit by accident — a subject line catches fire on a quiet inbox day, a segment happens to be unusually warm, a competitor goes down and traffic spills your way. None of that makes the campaign good. It makes it lucky.

A genuinely good campaign is not defined by the number it produced. It is defined by whether its underlying logic holds together — whether each part of it was built deliberately, in the right order, for the right reason.

That logic has five parts. Most underperforming campaigns are missing at least one of them, and it is almost never the part the team blames.

1. Intent — Why Does This Campaign Exist?

Before anything else, a good campaign has a reason to exist that is not "it is Tuesday and we have not sent anything this week."

Intent is the business or relationship problem the campaign is solving. It sits above the metric, above the offer, even above the audience. Every other decision in the campaign should trace back to it.

Weak intent sounds like: we should send something this week.

Strong intent sounds like one of these:

"This segment has not purchased in 45 days and is at risk of churning silently."
"We are launching a feature that solves the number one support complaint, and the people who complained need to hear about it first."
"Our list has grown 30% in two months and most of them have never received anything beyond the welcome email."

None of those mention a metric. They describe a situation — and a campaign built to address a real situation almost always performs better than one built to fill a send calendar, because every decision downstream has something real to optimise for.

The test: if you cannot state the intent in one sentence without using the word "engagement" or "content," the intent is probably not sharp enough yet.

2. Audience Targeting — Who Is This Actually For?

A list segment is not an audience. "Everyone who signed up in the last year" describes a database query. It does not describe a person, a situation, or a reason that person would care about what you are sending.

Good audience targeting answers a more specific question: what is this person's relationship with us right now, and what does that mean they need from this email?

That usually means targeting by behaviour and lifecycle stage rather than by static list membership. The difference between a broadcast and a targeted campaign is not the subject line. It is knowing what the recipient was doing before the email arrived.

Weak targeting	Behavioural targeting
"All subscribers"	"Subscribers who opened the last 3 emails but have not clicked"
"Customers"	"Customers who purchased once, 60+ days ago, no repeat"
"New signups"	"Signed up 7+ days ago, never opened the welcome series"
"Everyone on the promo list"	"Browsed the sale category twice in the last 5 days, no purchase"

The shift from a static list to a behavioural description does most of the work that copywriting and design get credited for. An email written for "everyone" has to be vague enough to apply to everyone. An email written for "browsed twice, did not buy" can be specific, because it is solving one situation, not ten.

3. Expected Behaviour — What Should This Person Actually Do?

Every good campaign is built around exactly one expected behaviour — the single action that would mean the email did its job.

This is the part most campaigns get wrong structurally. It is not that the action is unclear; it is that too many of them compete for attention. A typical underperforming email asks the reader to click through to a sale, also follow on social media, also forward to a friend, also reply with feedback, also download an app. Each additional ask dilutes the others to the point where none of them pull strongly enough.

A well-structured campaign can usually state its expected behaviour as a single sentence:

"Click through to the cart and complete checkout."
"Reply with which of the two options they prefer."
"Open the app and view the new feature."

Once that single behaviour is fixed, the rest of the email structure follows from it. The subject line creates curiosity about that one action. The body removes friction toward it. The call-to-action is the only button that matters. Anything that does not serve the one expected behaviour is a distraction the email does not need. This applies whether you are writing subject lines with AI or designing a manual campaign — the principle is the same.

4. Success Criteria — What Would Prove This Worked?

This is where most teams default back to a generic open or click number. But success criteria should be tied directly to the expected behaviour from step three, not pulled from an industry benchmark.

If the expected behaviour is "complete checkout," the success criterion is checkout completion, not open rate. If the expected behaviour is "reply with a preference," the success criterion is reply rate, not click rate — the email might not even need a link.

The discipline here is resisting the urge to track everything and call the campaign a success if any number looks decent. A campaign with a strong open rate but a 0.2% completion rate against an expected behaviour of "complete checkout" did not work — it just got opened.

A useful framing: decide the success criterion at the same time you decide the expected behaviour, before the email is even designed. If they are decided separately, the success criterion tends to drift toward whichever metric is easiest to make look good after the send. This is the same alignment problem that causes email teams to argue about performance — the standard was never set before the results arrived.

5. The Post-Send Learning Loop — What Does This Teach the Next Campaign?

A campaign that ends the moment it is sent is only half-built. The fifth part of the structure — and the one most consistently skipped — is feeding what happened back into how the next campaign gets planned.

This does not need to be complicated. It needs to happen on purpose, every time, with the same three questions:

Did the audience targeting match reality? Did the people in the segment actually behave the way the targeting assumed they would?
Did the expected behaviour happen, and if not, where did people drop off? Opened but did not click? Clicked but did not convert? The drop-off point tells you which part of the campaign to fix.
What would we change about the intent, targeting, or expected behaviour next time? Not the subject line — the structure underneath it.

Teams that build this loop in get measurably better at campaign design over time, because every send sharpens their model of what their specific audience responds to. Teams that skip it re-derive the same lessons from scratch, campaign after campaign, usually by repeating the same mistake until it becomes too expensive to ignore.

Putting the Five Parts Together

Here is the same campaign, built two ways — once without structure, once with it.

Without structure:

"Let us send a promo email this week to the list. Subject line testing 20% off vs. free shipping. We will check the open rate Monday."

With structure:

Intent: Cart abandonment has crept up 15% this month; this campaign exists to recover stalled checkouts before they go cold. Audience: Added an item to cart in the last 48 hours, no purchase, browsed the same category since. Expected behaviour: Return to the cart and complete checkout. Success criteria: Checkout completion rate from this segment, measured against this segment's historical baseline — not against open rate. Learning loop: If completion is low but click-through is high, the friction is at checkout, not in the email — feed that to the product team, not the next subject line test.

Both versions might use a similar subject line. The difference is that the second version has somewhere to go after it sends. It produces an answer, not just a result.

A Quick Self-Check Before Your Next Send

Run any upcoming campaign through these five questions before it goes out:

Can you state why this campaign exists in one sentence, without mentioning a metric?
Is the audience defined by what they have done, not just which list they are on?
Is there exactly one action this email is asking the reader to take?
Is the success criterion tied to that one action — not a generic open or click benchmark?
Do you have a plan for what you will do with the results, win or lose?

If any answer is "not really," that is the part of the campaign worth fixing — and it is rarely the part the post-send meeting ends up arguing about.

The Bottom Line

A good email campaign in 2026 is not the one that beats an industry benchmark. It is the one where every part — intent, audience, expected behaviour, success criteria, and the learning loop after it sends — was built deliberately and in the right order.

Benchmarks tell you how your number compares to someone else's number. Structure tells you whether your campaign was built to do anything in the first place. Only one of those actually makes the next campaign better.