Email marketing metrics that matter for an online shop

The metrics that matter for a shop#

A shop sends email to sell, so the metrics worth watching follow the path from inbox to order. Seven numbers cover that path and the health of your list. Each one answers a different question:

Outcome metrics tell you whether a campaign was worth sending. List health metrics tell you whether you can keep sending: inbox providers watch complaints and bounces closely, and a poor record affects your sender reputation for every future campaign.

Why open rate is no longer a reliable metric#

An open is recorded when a tiny invisible image in the email loads. Since 2021, Apple’s Mail Privacy Protection has stopped senders from using such pixels to learn about the reader, including whether and when the email was opened ^S1. Mail apps with the feature turned on load the content in the background, so many “opens” are recorded for people who never looked, while people who block images are never counted.

The result is an open rate that runs high and moves for reasons unrelated to your subject line. Even benchmark publishers add a warning that Apple’s changes may affect open-rate accuracy ^S3. Treat opens as a rough signal at most, and never use them alone to decide a test winner or to label someone inactive. Clicks require a real action, which is why click-through rate is the better engagement measure.

How to calculate each metric#

Use delivered emails as the base for most rates, since a bounced email had no chance to be read. Count each person once (unique clicks, not total clicks), and pick one attribution window for orders and keep it.

• Delivered = emails sent − bounces
• Bounce rate = bounces ÷ emails sent
• Click rate = people who clicked ÷ delivered
• Click-to-open rate = people who clicked ÷ people who opened (affected by inflated opens, so use with care)
• Conversion rate = orders attributed to the email ÷ delivered
• Revenue per recipient = attributed revenue ÷ delivered
• Unsubscribe rate = unsubscribes ÷ delivered
• Spam complaint rate = spam complaints ÷ delivered

A worked example

Ninefold, a small apparel brand, emails 2,000 subscribers about a new merino knitwear range. These are illustrative numbers, chosen to show the arithmetic:

Revenue per recipient is the number to compare across campaigns of different sizes. A VIP email to 300 people and a sale to the whole list can then be judged on the same scale.

Counting orders fairly

Conversion rate and revenue per recipient depend on which orders you credit to the email. A common rule is to count an order when the buyer clicked the email and then bought within a set number of days. A short window undercounts considered purchases such as a winter coat; a long one credits the email for orders that would have happened anyway. Whatever you choose, use the same window for every campaign, or the comparison between them means nothing.

Also decide whether discounts and returns come out of the revenue figure. A 20%-off campaign that brings in many orders can still earn less per recipient than a full-price new-arrivals email, and that is exactly the kind of difference these metrics should show you.

Healthy ranges set by inbox providers#

Most metrics have no official target; they depend on your products, prices and list. Spam complaints are the exception, because the largest inbox providers publish limits:

• Gmail: keep the spam rate reported in Postmaster Tools below 0.10% and avoid ever reaching 0.30% ^S2.
• Yahoo: bulk senders must keep their spam rate below 0.3% ^S4.
• Unsubscribes: both require bulk senders to support one-click unsubscribe for marketing messages ^{S2 S4}, and Yahoo asks that unsubscribes be honored within 2 days ^S4.

Neither provider publishes a bounce-rate limit. Treat any rise in hard bounces as a sign the list needs cleaning, and never email a hard-bounced address again; a suppression list handles that automatically.

One complaint in 1,000 delivered emails is 0.1%. On a list of a few thousand, three or four complaints can put a campaign over Gmail’s recommended level, so sending to people who asked to hear from you matters more than any subject line.

Email open rate benchmarks for ecommerce, and their limits#

Published benchmarks give a rough sense of scale. Mailchimp’s benchmark page, with data last updated in December 2023, lists an average for ecommerce of 29.81% opens, 1.74% clicks and 0.19% unsubscribes ^S3. It covers campaigns sent through Mailchimp with tracking on, to at least 1,000 subscribers, by users who reported their industry ^S3.

Read such figures with care. They come from one platform’s customers, they are averages across very different shops, the open rate is affected by Mail Privacy Protection ^S3, and each platform defines its rates slightly differently. A benchmark tells you whether a number is wildly off; it does not tell you whether your campaign was good.

The most useful benchmark is your own history. Compare each campaign with the last few you sent to a similar audience, using rates rather than raw counts.

How to read results after each campaign#

1. Check list health first. If complaints or bounces jumped, fix the audience before looking at anything else.
2. Look at click rate to judge the message: subject, offer and layout.
3. Look at conversion rate and revenue per recipient to judge the offer and the landing page.
4. If clicks were good but orders were not, the problem is usually after the click: price, stock, sizes or the product page.
5. Change one thing at a time in the next campaign, so you know what made the difference.

Smaller, better-targeted sends often improve every rate at once. See email segmentation for ecommerce and how to get more orders from email.

Seeing your results in Claude or ChatGPT with Tellnod#

With Tellnod connected to Claude or ChatGPT, you ask about a campaign in the same chat you sent it from. For example: “How did Thursday’s merino email do?” or “Compare it with our last new-arrivals email.”

The results card shows delivered, opens, clicks, unsubscribes and bounces, and, once your store is connected, orders within 7 days of a click. Campaigns are compared by rate, so lists of different sizes compare fairly. Opens are shown but should be read with the caution above.

Unsubscribe links and one-click unsubscribe headers are added to every campaign, and sending pauses automatically at a 0.1% complaint rate or a 2% bounce rate. Every send still goes through a confirmation card that only you can press. Plans start free for up to 500 contacts; see pricing, or add the connector from the connect page.

FAQ#