Impossibly low open rates on recent email

Hi,

Sugar Market users here.

Generally speaking our email open rates (over many many years) for our Members don't dip below 30%. That would be the absolute lowest I've seen. Since the introduction of Apple's email security features that push false positives (opens are registered by pre-loading the content) that number hasn't dipped below 30%. We have a very active and engaged membership.

Recently we sent an email and made a typo in the from address, listing the @domain.com part twice, accidentally. This led to a large uptick in soft bounces (rate was about 10%) and a dip in open rate. For the first time our open rate dropped to 12%. The same email in another language without this same domain error, had an open rate of 37%. A raw difference of 25%.

That open rate drop makes partial sense to me, as junk emails might not be pre-loaded the same way on Apple devices and are also much less prone to being legitimately opened. On the other hand, why did the open rate drop 25% if the bounce rate only increased 10%? Seems like something is missing. We'll return to this later...

Yesterday we re-sent the same email, fixing the from address, and leading with some apologetic text, "Due to an error the previous email may not have been delivered..." and so forth. That email's bounce rate dipped back down to normal rates (0.1%), but the open rate is stubbornly lagging at 22%. Note that the bounce rate dropped ~10% and the open rate increased ~10%!

There's a few reasons this seems impossibly low,

1) We know that our membership disproportionately uses Apple mobile devices. Apple's email security changes generated a noticeable increase in open rates that has sustained for several years.

2) We have many years of open rates, even before the Apple changes, and these two open rates are unprecedentedly low. We have data that supports this as recently as Friday May 25, 2023.

My hypotheses,

1) Members stopped using Apple devices or the relevant security features over the course of the last week. Very unlikely, but possible.

2) The emails are (soft bouncing) arriving in junk email inboxes, but not triggering the bounce metric in Sugar Market.

3) The emails are (hard bouncing) not arriving in inboxes at all, but are not triggering the bounce metric in Sugar Market.

I suspect that ~20% of this list is still hard or soft bouncing, but is not being counted as bounces. That would explain the lagging open rate.

Some things that might be relevant. (1) We do not have an unsubscribe link on these emails. We are required to send these to all these recipients, regardless of their opt-outs. It is a part of the by-laws that govern us. (2) We are in Canada, as are the majority of the recipients (unless they are traveling, for example).

I'm having trouble accessing my partner portal, but will be sharing the same details there.

Regards,
Ed

Parents
  • Anecdotally we have two of our Members confirming they are receiving our emails in their junk folders. These don't appear to be tracked as bounces in Sugar Market.

  • Hi Ed.  Sorry I am just getting to this now.  Have you had any further discoveries?  It would seem to me that the erroneous from address you started with may have caused more emails going to the junk folder?  I don't believe an email going to the junk folder is the same as a bounced email, hence the different in the metrics you are seeing.

  • Yea, we have a widespread deliverability issue. I did some Excel and am generating open rate reports by domain to use as a proxy for deliverability. This might oddly be the upside of Apple's email privacy stuff...makes it clear when an open rate is too low.

    Plan now is to take existing campaigns and roll them out in progressively higher quantities over the course of this week, like a miniature IP warming exercise. Fortunately, this involves using new content, sending address, and much lower volume. Phase 1 of this today appears to be going well. Open rate on gmail.com was 1% last week and is 18% after 2.5 hours of data accumulation today, so we think we are on the right path.

    I think the deliverability should log as a soft bounce (spam-related is the bounce type we usually see in this scenario), that happens sometimes, but maybe not all the time. I don't really know, I suppose.

    I think if I was going to make a feature request based on this experience, it would be something that causes an error on the setup screen in email builder 2.0 if the from or reply-to address fields have an '@' symbol. Obviously typos are our fault and those will happen sometimes, but in this case it's also an invalid email format so maybe it can be prevented at that stage in future. Again, it's my error that caused this, but that would be a potentially helpful tool change to prevent the issue for your other clients as well as us in the future.

Reply
  • Yea, we have a widespread deliverability issue. I did some Excel and am generating open rate reports by domain to use as a proxy for deliverability. This might oddly be the upside of Apple's email privacy stuff...makes it clear when an open rate is too low.

    Plan now is to take existing campaigns and roll them out in progressively higher quantities over the course of this week, like a miniature IP warming exercise. Fortunately, this involves using new content, sending address, and much lower volume. Phase 1 of this today appears to be going well. Open rate on gmail.com was 1% last week and is 18% after 2.5 hours of data accumulation today, so we think we are on the right path.

    I think the deliverability should log as a soft bounce (spam-related is the bounce type we usually see in this scenario), that happens sometimes, but maybe not all the time. I don't really know, I suppose.

    I think if I was going to make a feature request based on this experience, it would be something that causes an error on the setup screen in email builder 2.0 if the from or reply-to address fields have an '@' symbol. Obviously typos are our fault and those will happen sometimes, but in this case it's also an invalid email format so maybe it can be prevented at that stage in future. Again, it's my error that caused this, but that would be a potentially helpful tool change to prevent the issue for your other clients as well as us in the future.

Children
No Data