Email open rate is unique opens divided by emails delivered — 3,920 opens from 9,800 delivered is 40 per cent. Since Apple Mail Privacy Protection began pre-loading tracking images for a large share of recipients, that figure has been inflated by an unknown and shifting amount, because a pre-load registers as an open whether or not anyone read anything.
Click rate did not change. A click is a real action a person took, which is why it is now the number worth optimising toward.
What should the denominator be?
Delivered, not sent. Bounces have to come out or every rate is understated, and the gap between the two is the first thing to check when a rate looks wrong.
| Metric | Formula |
|---|---|
| Delivery rate | Delivered ÷ sent |
| Open rate | Unique opens ÷ delivered |
| Click rate | Unique clicks ÷ delivered |
| Click-to-open rate | Unique clicks ÷ unique opens |
Click-to-open is the one most damaged by the privacy change, because its denominator is the inflated figure. A falling click-to-open rate on a list that has gained Apple Mail users is measuring the mail client mix rather than the campaign.
Why unique rather than total?
Because total clicks double-count one enthusiastic reader clicking three links, which makes a campaign look broader than it was. Unique clicks count people; total clicks count events, and they answer different questions.
Both are worth having. Unique tells you how many people acted; total divided by unique tells you how thoroughly the ones who acted engaged, which is a useful signal about content length and link placement.
What is the ratio worth watching?
Click rate against unsubscribe rate. A campaign that lifts clicks while lifting unsubscribes proportionally is borrowing from the list rather than earning from it — and the list is a finite asset.
Two hundred and forty-five clicks from 4,900 delivered is a 5 per cent click rate. If the same send produced 60 unsubscribes, that is 1.2 per cent, and whether the trade was worth it depends on what the clicks were worth. This is the same break-even thinking as an ad campaign, applied to an asset you own rather than one you rent.
Deliverability sits underneath all of it and is measured before any of these rates. A delivery rate below about 95 per cent is a signal to stop optimising content entirely and look at authentication, list hygiene and complaint rate — because a campaign that does not arrive cannot be tested.
Where does a list plateau?
Where churn equals signups, and most projections miss it. Signups tend to be roughly constant, driven by traffic and offer, while churn is a percentage and therefore grows with the list.
A 5,000 list gaining 350 and losing 100 a month is growing 5 per cent monthly and reaches about 8,900 in a year. But the 100 lost was 2 per cent of 5,000; at 17,500 the same 2 per cent is 350, which exactly cancels the signups. That is the equilibrium, and it arrives whether or not anyone planned for it.
Moving the plateau means either raising signups or lowering churn, and lowering churn is usually cheaper. A one-point reduction in monthly churn moves the equilibrium substantially, because it is dividing into a constant.
Two further metrics survived the change and are worth watching in its place. Revenue per send divides the campaign revenue by delivered emails and is the closest thing to a single measure of a send being worth making. Reply rate, on lists where replies are plausible, is the strongest engagement signal there is and the hardest to fake.
What should you optimise now?
Clicks, revenue per send, and list health, in that order. Subject-line testing against open rate is testing against a number that partly measures mail clients, and the conclusions it produces are unreliable in a way that is invisible.
Where open rate is still useful is as a relative signal within one segment over time, provided the client mix is stable. As an absolute figure or a cross-campaign comparison it has stopped being trustworthy, and no amount of care in the calculation fixes that.
Questions people ask
What is a good click rate? It depends on the list and the offer far more than on the industry, and the only comparison worth making is against your own recent sends to the same segment.
Should I stop reporting open rate? Report it with the caveat rather than dropping it. It still detects deliverability collapses, which is a real use.
Does the privacy change affect all recipients? Only those using the affected mail clients, and the share differs by list. That is precisely why the inflation is unknown rather than a fixed correction you could apply.
Is a 40 per cent open rate good? It was, before the change. Now it depends on a client mix you cannot see, which is the whole problem.
How do I reduce churn? Send less to people who do not engage, and segment rather than blasting. Most churn is a relevance problem presenting as a frequency problem.
Does list size affect the rates? Larger lists usually show lower rates, because they contain more people acquired further from the offer. Comparing a 50,000 list against a 2,000 one on rate alone compares two different audiences.
Judge sends on clicks and the list on its plateau. The email open rate calculator and email click rate calculator both use delivered as the denominator, and the newsletter growth calculator shows where signups and churn balance out.