SPF, DKIM, & DMARC are all fine. I'm using Amazon SES. Their spam filters are awful and are harming an important project. They're permanently (I waited over a month) sending all emails from a shared Amazon SES IP range to spam because 0.3% of users who received an email from that IP marked it as spam instead of unsubscribing... I can't believe they think that is a reasonable thing to do... People were taught for years to mark emails as spam instead of unsubscribing since unsubscribing can confirm your email to spammers. 1% of people may also simply be dumb, or may be sabotaging your business on purpose. Making such a drastic decision based on less than half of 1% of email recipients is ridiculous. Allowing the 1% lowest common denominator to determine what the other 99% are allowed to see is such a horrible policy. I've done everything in my power to highlight the option to unsubscribe instead of mark an email as spam, including the ridiculous step of opening the email with "You're receiving this email because of xyz. Please blocklist your email if you don't want any more" instead of "Hi, here are the details of an important update". For years, I've been responding to people who applied via a google form on our website (1.2+ million at this point). When simply sending a first email telling them "you did or did not qualify", the Amazon SES complaint rate has always averaged between 0.3% and 0.5%, and the Google Postmaster rate always jumps between 0%-2%. There's nothing that can be done to bring the rates down besides sending people unnecessary emails in order to lower the average. So sending people actual spam is the only way to decrease the spam rate below that extremely low level... There are times when 50% of my incoming [gmail] emails go to spam. When I contacted Outlook about this their response went to my Outlook junk folder. Their spam filters are horrible and they don't seem to care at all. And it's causing major problems with a majority of people missing important emails. Eg: https://arstechnica.com/cars/2024/02/a-spam-folder-may-have-foiled-andretti-cadillacs-f1-entry/ Yahoo said I should request a better IP pool from Amazon SES. I said: I searched Stack Overflow and Hacker News and found loads of other people with the same problem. I found a way to submit bulk mail appeals to Gmail, Yahoo, Outlook, etc. I think my appeal to Gmail worked (had to wait a few weeks), but Outlook and Yahoo basically just told me "Too bad, your IP pool is bad". This is such a stunningly bad policy. There has to be a way to protest it and get it changed.
It's disheartening to see your important work affected by spam filters that seem to penalize you for things beyond your control. Emailing is meant to be a simple process, but it feels unfair when algorithms overshadow genuine communication. It might be worth rallying support from others facing the same issue. sometimes, collective voices can push for change.
I agree completely. I've been searching the internet for communities that discuss this subject. If you know of any, please let me know. I found: * Hacker News news.ycombinator.com * Server Fault on Stack Exchange https://serverfault.com/ * This forum. * Mautic https://forum.mautic.org/t/using-amazon-ses-mail-landing-in-spam-folder/26994/8 * Warrior forum: https://www.warriorforum.com/email-marketing/ - Unfortunately, new members are required to have 5 posts before they can start their own thread, and the existing threads are old & closed. This blog lists some ways to contact the major provider's bulk email departments: https://www.emailtooltester.com/en/blog/why-are-my-emails-going-to-spam/
They definitely need to make it easier but here's a wee anecdote from my early days We had software that people could download for a free trial and they'd leave their email address. Occasionally customers would email with questions or for support. A small group would get shitty that we didn't reply. They were all AOL customers - we ended up having to put a note on our website about AOL. I downloaded one of those ping tools that showed you the nodes that your ping passes through and I could see exactly where the pings ended if we were sending them to AOL. BUT Have you ever tried tracking down who is responsible for one of these nodes? Back then, 25 years ago, it was impossible. I documented it & sent all the evidence to AOL and they replied with IDGAF even though it was impacting their customers (other US email addresses were ok). And then one day, years later, someone rebooted or replaced that machine and the emails started flowing.
I found a forum where the main topic is email - emaildiscussions.com. I ran an experiment: Who is marking legitimate emails as spam? N = 1.2 million experiment https://www.emaildiscussions.com/showthread.php?t=80728
It looks like someone on that forum didn't like what I wrote and deleted it. I'm copying it here because I need to share it with the major email providers so they can understand how problematic and erroneous their spam filters are. And then hopefully fix them. People continue to contact me complaining that they've never heard back, when in fact they were sent multiple important updates that landed in their spam. It is severely harming my project. Background: I have a Google Form on my website that has collected more than 1.2 million applications over the past 4 years. People are applying for contract work that pays well if you are healthy. The vast majority of people who apply are not healthy. My historic complaint rate has been pretty consistent: 0.3 to 0.5% in Amazon SES, and 0 to 2% in Google Postmaster. I noticed that the vast majority of complaints came from people who were sent an email saying they weren't healthy enough. I thought they were marking the emails as spam because they were unhappy about the result. However, recently it's become so bad that despite people constantly asking for updates, I can't even send out neutral updates without all my emails going to spam. I've been reading about other people's experiences and trying to figure out what I can do to solve this. I read a forum owner describe how they had their Amazon SES account temporarily suspended due to a surge of spam complaints. It turned out that they all originated from a single user who moved a bunch of old forum notifications to spam in one go. When they messaged the person about it on the forum they apologized and said they were just trying to "mute notifications". This is a great example of two major flaws with spam filters: There's a minority of users who use/abuse spam/junk folders as an easy way to mute/unsubscribe and manage their inbox. They are not intentionally reporting the emails as SPAM and they probably do not understand the consequences of their actions. There needs to be changes to the algorithms that Gmail, Outlook, and Yahoo use to account for these types of users and decrease the impact of their actions. A bunch of emails moved to spam in one go by this one person should not have resulted in a ton of individual spam complaints which terminated the sender's SES account. There needs to be adjustments to the spam algorithms to account for this. My experiment: Previously, I discussed how terrible of a policy it is to cater to the lowest 1% common denominator. This plus the fact that even neutral emails (vs "bad news", "you didn't qualify") are being marked as spam made me consider that it's the 1% most cognitively deficient people who are marking emails as spam. Given the unique nature of what people are applying for, I'm actually able to test this hypothesis. Some background information: Obesity and diabetes are rampant. Both conditions are associated with cognitive impairment. So the most unhealthy people are severely cognitively impaired. pmc.ncbi.nlm.nih.gov/articles/PMC2769828/ https://www.nature.com/articles/s41574-018-0048-7 pmc.ncbi.nlm.nih.gov/articles/PMC10011899/ pmc.ncbi.nlm.nih.gov/articles/PMC5528145/ pmc.ncbi.nlm.nih.gov/articles/PMC10615478/ https://github.com/MaximilianKohler/Archive/wiki/Health#obesity-and-brain-function So instead of sending out the neutral email updates to all applicants, I sent them out to only the ~20% healthiest applicants. The result: The result was minimal spam complaints, and my open rates gradually increased from ~15% to 75% as my emails stopped going to spam. When emailing the 20% healthiest applicants, the highest spam rate for Google postmaster was 0.6, but it was usually 0 to 0.3. A huge decrease from the usual 2% max, 1% average. AOL, Outlook, and Yahoo were still sending my emails to spam. It's ridiculous that they're even worse than Gmail given that there are times when a majority of my incoming, legitimate Gmail emails are marked as spam. Perhaps those other providers have more users who abuse/use their spam/junk folder as a way to unsubscribe and manage their emails, or perhaps those providers' algorithms are even worse than Google's. Postmaster was completely missing the data for when I started emailing the unhealthy portion, but using my SES metrics, the complaint rate dropped to 1/3rd when emailing the healthy people, then doubled when I emailed the unhealthy ones. However, when switching to emailing the unhealthier 80%, there was no obvious "massive increase in spam complaints" resulting in my emails going to spam again, or open rates drastically decreasing. Therefore, it seems that the overall spam rate decreases as I email larger numbers of people. Eg: emailing hundreds of thousands, vs emailing a few thousand. That's likely a problem with how they calculate their rates. And this is likely contributing to the higher spam rate problem when I'm emailing fewer people. Thus, the "solution" seems to be: Start off emailing the 20% healthiest people, then you can email everyone without issue. A month later my Postmaster data showed up and it didn't go over 0.2%. So there was no major increase in complaints, which tells me that a huge part of the issue is the way their algorithm works. If you're emailing a few thousand people per day you get penalized harder than if you are emailing tens-hundreds of thousands. Thoughts: I could probably test this further by removing all applicants who have diabetes, and so on... But I think this makes spam filter deficiencies clear enough, and they need to be fixed by Google, Outlook, and Yahoo. It is absolutely ridiculous that I should have to segment my subscribers like this in order to avoid all my emails going to spam. Currently, if more than 3 out of 1000 people mark your email as spam, the other 997 will likely never see it. This is completely absurd and makes it extremely easy to sabotage businesses/projects. For example: If you're working on something controversial If you're working on something that is a threat to monied interests If people make misleading social media content about you If you send people bad news If you word things poorly, or if people don't like your wording or find it suspicious You're much more likely to go over that complaint rate and have all your emails sent to spam. None of these should have any relevance to whether or not your emails land in spam. But right now they do. And the major email providers don't care. All of this would matter much less if most people were regularly checking their spam/junk folders, but it seems that the vast majority do not. Email providers need to account for this. DMARC: Another major issue caused by these faulty spam filters is that people will be more susceptible to phishing attacks since they have to whitelist our domain. This makes a "quarantine" policy meaningless. Thus, I changed it to "reject" because I received some phishing emails from someone who spoofed my domain. However, there were some emails sent via Amazon SES that failed SPF and thus were rejected. It's not possible to keep track of all the failed emails and resend them, so a "reject" policy seems nonviable. Thus, I had to switch back to "quarantine", and people who have whitelisted our domain will be more susceptible to phishing attacks.