Email marketing metrics explained: which ones matter

The short answer

The email marketing metrics that matter most are delivery, clicks, unsubscribes and the business result, such as enquiries or orders. Open rate is unreliable because some mail apps load emails automatically, which can record an open nobody made. Judge a campaign against the goal you set before sending and against your own earlier sends, not against other businesses.

Email marketing metrics grouped by the question they answer

Email marketing metrics are the counts and percentages your email tool reports after you send. The useful way to read them is by question: did the email arrive, did people want it, did it get attention, did it prompt action, and did that action pay. Each question has a metric that comes closest, and each metric has a blind spot.

Move down the table as you gain experience. A new sender should watch the first two rows closely. Once the basics are healthy, the last row is where decisions come from.

Delivery metrics: what arrived and what bounced

A message counted as delivered was accepted by the receiving mail server. That is not the same as reaching the inbox. A server can accept a message and still file it as spam, and your tool usually cannot see that.

Bounces come in two kinds. A hard bounce means the address does not exist or the server has refused you permanently, so remove it. A soft bounce is a temporary problem such as a full mailbox, and a single one rarely matters. If the same address soft bounces across several sends, treat it like a hard bounce.

Spam complaints are usually the smallest number and one of the most serious. They happen when a recipient presses the report button instead of unsubscribing. If complaints rise, look at your sign-up process and your expectation-setting, and not just at the email itself. Set up the sender authentication your email provider asks for, since it affects how mailbox providers treat your messages.

Why open rate is unreliable

Open rate counts how many emails your tool recorded as opened, divided by the emails delivered or sent depending on the tool. The recording works through a tiny invisible image that loads when the email is displayed. That method has several weaknesses.

Apple documents a setting called Mail Privacy Protection in its Mail app. When it is on, the app loads remote content privately in the background when a message arrives, not when the person views it. Your tool may then record an open for someone who never looked at the email. People can switch the setting off, so the effect varies from list to list. Check Apple's current documentation for the exact behaviour.

The error also runs the other way. Some mail apps block images by default until the reader allows them, so a person who reads the whole email without loading images is not counted. Some security software may fetch content automatically before a person sees the message.

The result is a number that can be inflated for one reader, understated for another, and different on every list. Use it for one limited purpose: comparing two subject lines sent to similar halves of the same list, and even then treat a small gap as noise. Do not use it as a headline figure, and do not set targets against it.

Open rate vs click rate and click-to-open rate

Click rate is the share of recipients who clicked at least one link. Most tools divide the number of people who clicked by the number delivered, though some divide by the number sent. If two hundred emails are delivered and ten people click, the click rate is five per cent of those delivered. This is an arithmetic illustration, not a benchmark.

Click-to-open rate divides clicks by opens. It was designed to show how persuasive the content was once someone opened. Because it uses opens, it inherits the open rate problem. Read it only as a rough comparison between sends of a similar kind.

Always find the definition in your tool's help pages before comparing two reports. Two tools can show different click rates for the same send because their formulas differ, and a tool can change its definition. Counting unique clickers or total clicks also gives different results, since one keen reader can click five times.

Clicks are a better signal than opens because they need a deliberate action. They still have limits. A click shows interest in the link, not that the page delivered. For that, look at what happened on your website afterwards.

Which metric to watch for each kind of email

If you publish a regular newsletter, the guide to starting a newsletter explains how to give each issue one job, which makes the right metric easy to pick.

Connect email to results on your website

The email tool stops at the click. To see what visitors did next, tag the links in each campaign with campaign parameters so your analytics tool can label the visit as coming from email. Use the same naming pattern every time, such as the month and a short name for the campaign, so reports stay tidy.

Then count what matters to the business: orders, bookings, enquiries or downloads. Add a line to the order or enquiry form asking how the person heard about you, and keep a simple log. Analytics will under-count some email visits, for example when a reader forwards a message or opens a link in a private window, so the log is a useful second source.

Email often helps without getting the credit. A customer may read three emails, search for your name a week later and buy. Last-click reports credit the search. Do not judge email only on direct sales if your customers tend to take time before buying.

How to judge whether a campaign worked

A campaign worked if it did what you set out to do, measured against your own history. Follow the steps below.

1. Write the goal before sending: for example, a set number of enquiries from the new service page, or more bookings for a quiet week.
2. Choose one main metric that matches the goal, and one or two supporting ones.
3. Compare with your last few sends of the same kind to a similar audience. A single send tells you little.
4. Check the cost side: unsubscribes, complaints and replies that asked to stop.
5. Write down one change for the next send, and change only that.

Small lists swing widely, because a handful of people can move a percentage by several points. Look at the counts and the trend over several sends, not just the percentage. If a campaign missed its goal, check the offer before the subject line. A weak promise cannot be fixed with a clever heading, and the guide to writing a value proposition helps you test whether the offer is clear. Noting the goal of each send in your content calendar makes this review quick.

Benchmarks, list quality and list growth

Published average rates come from lists with different audiences, sending habits and mail-app mixes. They tell you little about your list. Your own last six months is a better yardstick.

Track list growth alongside campaign results: new subscribers, where they came from, and how many left. A list that grows from a clear offer on your site behaves differently from one built by bulk-adding contacts. The guide to email list building for a small business covers sources worth keeping.

Email data also feeds other channels. Some advertising platforms let you upload a customer list to build an audience, which Meta lists as one source for custom audiences. Only do this if the people on the list gave permission that covers it, and read the platform's terms first. The guide to how Meta ads work explains where that fits.

Mistakes that make email numbers misleading

• Chasing open rate and rewriting subject lines when the real problem is a weak offer.
• Judging a campaign by one send, or by one lucky or unlucky week.
• Comparing a newsletter to a promotion, or a list of customers to a list of prospects.
• Ignoring unsubscribes because they look small, when they are often the clearest feedback on frequency.
• Changing the subject line, the offer and the send time together, then not knowing which one mattered.

Choose three to five numbers, record them after every send and look at them monthly. If you want help setting up tracking and reporting across channels, the digital marketing service page describes what Liftup Technologies offers.

Related guides

• Digital marketing
• Newsletter how to start one
• Social media content calendar how to make one
• Value proposition how to write one
• Email marketing vs social media marketing
• Marketing funnel explained

Questions and answers

Which email marketing metrics matter most?

Clicks, replies, unsubscribes and the business result, such as orders or enquiries, matter most. Delivery and bounce numbers matter while you are getting set up, because they show whether messages reach people at all. Treat opens as a rough signal only.

Why is open rate unreliable?

Opens are recorded when a hidden image loads. Some mail apps, including Apple's Mail with Mail Privacy Protection turned on, load that content in the background, so an open can be recorded without anyone reading the email. Other apps block images, so real opens can be missed.

What is the difference between open rate and click rate?

Open rate estimates how many people displayed the email. Click rate shows how many clicked at least one link. Clicks need a deliberate action, so they are a firmer signal of interest, although the exact formula varies between email tools.

How do I judge whether an email campaign worked?

Write down the goal before you send, pick the one metric that matches it and compare it with your own recent sends of the same kind. Check unsubscribes and complaints as the cost side. One send is weak evidence, so look at the trend across several.

Should I compare my email results with industry averages?

Use averages with great care. They mix lists with different audiences, sending habits and device mixes, and the way each tool measures opens and clicks differs. Your own history over the last few months is a more honest comparison.

What is a good click rate for a small business?

No single figure applies to every list, because the offer, audience and measurement method all change the result. Build your own baseline from several sends and aim to improve on it. Treat any number quoted without that context as a guess.