Email engagement has two instruments: an image that loads, and a link that is followed. Open rate rests on the first. Click-through rate rests on the second. The second held up when privacy features broke the first.
Which denominator?
Click-through rate divides recorded clicks by delivered messages. Click-to-open rate divides recorded clicks by recorded opens. State whether the numerator counts all requests or unique recipients.
| Rate | Denominator | What it answers |
|---|---|---|
| Click-through rate | Messages delivered | Recorded clicks relative to delivered messages |
| Click-to-open rate | Messages that registered an open | Recorded clicks relative to recorded opens |
Compare the rates using the same click numerator and scope. Click-to-open is larger only when its nonzero open denominator is smaller than the delivery denominator. For example, 20 clicks over 100 deliveries is 20%; those clicks over 40 recorded opens is 50%. Neither ratio measures how many people read or acted.
An event-time report can count opens for messages delivered outside the reporting window. The stats API uses that attribution, so its windowed open count need not be a subset of its delivered count. A zero denominator leaves the ratio undefined. Automatic image fetches can also change the open denominator without a change in reading.
A benchmark taken from one and applied to the other says nothing. Ask which denominator before you compare your rate to anyone else's.
How is a click measured?
Through a redirect, which is why the metric held up.
Bird rewrites the links in your HTML to point at a tracking hostname. The recipient's browser requests the rewritten URL, an email.clicked event is recorded with the original destination, and the browser continues to that destination. The open and click tracking guide has the shape.
That hostname sits on your own sending domain rather than on a shared one. A recipient hovering a link sees your domain. The link reputation accrues to you rather than to a pool. Nothing is rewritten until the record verifies, which tracking domain covers.
A redirect needs no image and no HTML rendering. Someone who reads in plain text, blocks remote images, or uses a text-only client still produces a click when they act.
Why do I have more clicks than opens?
Because the two instruments fail in different places.
An open needs the pixel to load. A client that blocks remote images never loads it. The plain-text part of a message carries no pixel at all. A rewritten link still works in both cases, so the click lands where the open does not.
Forwarding produces the same shape from another cause. A forwarded copy carries the same rewritten links, so a click can arrive from someone who was never on your recipient list. No open of theirs was ever yours to measure.
So clicks above opens is not a defect. One population is reading with images off. Another is being forwarded your mail. Open and click tracking documents both causes.
Is a click proof someone acted?
No. Security tooling follows links. Each fetch looks like a click.
Filters, gateways and anti-abuse systems fetch URLs to see where they lead. Spamhaus, which publishes email blocklists, states its own practice and says it is not alone in it:
Please note – Spamhaus reserves the right to engage with the spam we receive and even to follow non-subscription links. There are many anti-spam appliances that now do this also. The fact that an email is opened or a link is clicked does not mean consent has been given.
So a click is not permission. Treating one as an opt-in is a consent error rather than a measurement error.
Does any of this apply to SMS?
The instrument does. The assumptions around it do not.
An SMS click is measured the same way, through a rewritten link. SMS has no open to record. The message is short enough that the link is usually the whole call to action, so the rate is not diluted by other content. And the recipient sees the link before tapping, which makes a recognizable hostname on your own domain more consequential than it is in email.
An SMS click rate and an email click-through rate come from different message lengths, different consent regimes and different expectations. A ratio between the two describes those formats rather than your program.
What should I do with the number?
Use recorded clicks to investigate a journey, then measure the intended action in your application.
A click can come from an anti-abuse system or a forwarded message, so it cannot identify the recipient's action on its own. For a purchase campaign, compare completed purchases for the same audience and period. For account setup, compare completed setup.
Keep opens, clicks and completed actions separate. Check unsubscribe rate and complaint rate alongside them, because more link requests do not establish that fewer people want to leave.
In short
The denominator determines which rate you report.
Both ratios use recorded clicks. Compare them with the same numerator and scope; a zero denominator leaves either ratio undefined.
A click needs no image.
Link tracking works through a redirect rather than a pixel, so image blocking and privacy proxies do not distort it the way they distort opens.
Clicks above opens is a normal reading, not a bug.
A recipient with images off can click without registering an open. A forwarded copy carries your links to someone who was never on your list.
A click is still not consent or attention.
Anti-abuse systems follow links. Spamhaus says so about its own practice.