There’s something about bounces

I’ve shared a version of this image repeatedly. I think it was only my Facebook friends that got the stick figure screaming in frustration, though.

An overlapping circles diagram showing bounce classifications from 3 different ESPs and how they overlap (and don't) with each other and with my classifications of the underlying error.

The reality is bounce handling is one of the most frustrating pieces of email delivery. Not only that, many people in the email space treat it as a simple process. It’s really not as simple as we’d like it to be.

The above image was created based on docs from 3 different ESPs a client was using. They wanted to normalise their bounce handling across ESPs, and asked me for policy recommendations. I ended up digging through a bunch of docs from their 3 ESPs. I recorded the reasons as reported in the docs in a colored block corresponding to the ESP, then dropped them in the appropriate circle: soft, block or hard.

The shaded circles are based on my interpretations of why these bounces happen.

  • The big grey circle surrounds bounces due to reputation issues.
  • The green circle is primarily networking and technical issues.
  • The top purple circle is non existent or bad addresses

Note, nothing here indicates how we should react to the bounces, this is just a categorisation activity. This classification also has nothing to do with what the actual SMTP response is.

Just remember, next time someone says bounce handling is simple: they’re wrong.

Related Posts

Thoughts on bounce handling

This week’s Wednesday question comes from D.

What are your thoughts on bounce handling

Read More

Increase in bounces at Y!

I’ve been seeing reports over the last few days about an increase in bounces at Yahoo. Reliable people are telling me they’re seeing some increase in “invalid user” bounces.
You may remember Yahoo announced an overhaul of their mail product back in December. Reliable sources tell me that this is more than just interface revamp. In the back end, Yahoo! is removing older products with few users and security problems. This fits in with the changes CEO Mayer has been making with the company: slim down and stop supporting unprofitable products.
It makes sense that while engineers are looking at the guts of the email program and cleaning up the cruft, they will also disable long unused email addresses. This will result in higher unknown users for some senders.
What’s interesting to me is that the reports are somewhat sporadic. Some senders are seeing a huge percentage of bounces, some are seeing the normal percentage. I expect this difference isn’t anything more than how actively a sender purges based on engagement. Senders that purge unengaged addresses are going to have already removed a lot of the addresses Yahoo! is now purging from their database. Senders that keep sending to their whole list, are going to see a lot of unknown user bounces.
I’ve asked a few folks and people who’ve responded told me that spot checks showed all the addresses turning up as invalid had no engagement for long periods of time.
If you are seeing a lot of bounces at Yahoo! over the last few days, you need to remove those addresses from your lists. I also recommend looking at the engagement statistics of these newly purged recipients. This will tell you, approximately, what an abandoned address profile looks like. You can use that information to make good decisions about purging unengaged users at other ISPs as well. Not only does this lower costs, because you’ll be sending to less non-responsive email addresses, it will also improve delivery at many ISPs.

Read More

Share your average bounce rates

The question came up on slack this morning about bounce rate benchmarks. What are the normal / average bounces that different ESPs see? Does region matter? What’s acceptable for bounce rates?

Read More