We got a great five-letter domain name for our web app, but we're inheriting a legacy of spam being sent from that domain. And hence, Gmail immediately dumps all transactional emails into spam. This is just a tiny problem.
We're using SendGrid to handle our transactional emails and have them white-labeled to our domain. We're not blacklisted by any services.
What can I do to overcome this spam problem?
I'm not sure there really is that much you can do, other than making sure that your dedicated IP (Which SendGrid has probably provided for you) is continually building reputation.
One of the things I'm trying to figure out is if your domain gets flagged for spam once, is it possible to recover? I am hoping it is.
Related
My team supports a website for a client, and we use SendGrid to send email related to the site on their behalf.
We do not have anything to do with their own email server and I don't at present know anything about it.
So far as I can work out, SendGrid has proper authentication and is an authorised sender for their domain, and almost 98% of email is delivered successfully.
However, we have had a handful of bounces with the reason "550 relaying denied" and all of these were to addresses at our client's domain (the same one as their website and the from address of the emails.)
Most emails to their domain were delivered successfully.
Unfortunately I don't have access to the full headers of the bounce emails, only the reason.
I understand that in general this error can either be caused by
the sender not being authenticated correctly. I am very far from being an expert in this but so far as I can tell, there is nothing wrong there. Or
a DNS or similar misconfiguration on the part of the recipient's email domain. I have even less understanding about this and I have no access or responsibility for the client's email server.
My main question is, is there any way the domain being the same as the from address could be related? Being as the email is claiming to be from the same place it's sent to, is it possible for that to affect how it's handled by relays?
If not, I'd also appreciate any pointers on where to look for the issue (or what to advise the client to look at if the problem is likely to be from their end.) I have been trying to research issues with email configuration and authentication but I am very much a novice in this area.
Thanks in advance.
The domain being the same could very well be related, but normally when that happens, the receiving server refuses all mail purporting to be from itself.
Separate from DKIM & SPF, most mail servers believe they alone are responsible for the mail from their domain.com. As such, a lot of them have anti-phishing filters that reject "outside" mail that claims to be from themselves. It's like "You can't be Carrie, I'm Carrie! Go Away!"
The fact that it's only some mail is interesting. The error being relay denied may also be key, though these anti-phishing filters often use "fake" errors to not give away the game.
Do the recipients of the messages that are being rejected have some kind of internal forwarding applied? That may be the cause, in which case that bounce reason is honest.
Or they may have a more defined anti-phishing feature, only rejecting mail From or For certain addresses. You can try testing certain combinations, and see if anything is repeatable.
Ultimately however, it will come down to working with the receiving mail domain's admin, and either updating those rules, or whitelisting the SendGrid IPs that are sending the mail to them.
We have a powerful VPS currently having various websites. This websites although they do not spam have had their i.p. emails blacklisted in the past? we keep fighting against getting the i.p.. delisted because it otherwise affects all websites email deliverability. It seems too vulnerable that if the I.P. gets blacklisted then everybody loses business not being able to contact their clients. I know large websites have strategies to avoid this, not sure how they do it. I would like to know an experts advice on how to deal with this problem. In summary what are recommended best practices for business email deliverability when depending on one I.P.? or is there such thing as dynamic I.P.s? Open to any options to solve this critical problem for us.
There are many steps you can take to insure that your IP does not get blacklisted in RBL's.
You can assign a unique IP for each website and you can assign same unique IP for outgoing mails so all the domains will not be affected if IP gets blacklisted.
Make sure you have enabled SMTP authentication for all the domains.
You can set SPF records for all the websites. You can create an SPF record from below URL.
http://www.spfwizard.net/
Create Domainkey (DKIM)
Use strong password for email accounts as weak password are more prone to get compromised.
Hope this helps.
I am trying to send mails with mailgun. My DNS config (SPF,DKIM) seems to be ok and are being validated in mailgun service. I can send mail to several users with gmail, live and most others mail providers. However, I have a problem when I sent an email for email accounts of my university.
The message is rejected with the following alert:
"554 5.7.1 : Client host rejected: MX-CIDR"
My current DNS settings are:
TXT # "v=spf1 include:mailgun.org ~all"
MX 10 mxa.mailgun.org.
MX 10 mxb.mailgun.org.
DKIM was validated as well. I checked my domain at mxtoolbox and the dns config pass in all tests. I did not find errors related with that alert in others questions. May someone help me to fix it?
Update 1:
Just some more informations:
1) I dont send, and I have absolutely no intention to send spam. I created an educational website, used by students and instructors, and they send messages sometimes between each others. I also send mail to confirm registers, recovery password, as a lot of others websites do. I only send messages to people who was agreed with my terms of service, that includes the information about my mail policy. It is a small service, I never sent more than 2,000 messages in a month (I have 800 registered users so far)
2) I do not believe I was blacklisted, mxtools verify several blacklists databases and my IP have passed in all verifications. Also, the server is not rejecting all messages from my IP, I can send messages with my personal email with the same domain, but I use different services to handle my personal inbox with my domain and the emails send by my website. So, I guess it may be a DNS record mistake.
3) I only use mailgun (or others transactional email services like mandrill or sendgrid) because it is highly recommended (and easy). I use a small VPS and it is hard to configure my own email service (I am a programmer, I am not an expert in that kind of configuration). If exists negative factors about the use of these systems, I really like to know and learn more.
I see no evidence posted that the reason the receiving mail server is rejecting your mail is because of your SPF records.
There isn't even any evidence here that the receiving mail servers are even performing SPF checks on their incoming mail.
Can you explain why exactly you believe that this has anything to do with SPF?
Just because someone's rejecting your mail, and you happen to be messing around with your SPF records, doesn't mean that the reason for your mail being rejected is due to your SPF records.
The only ones who can tell you exactly why your email is being rejected, and what needs to be done to fix it, is the receiving mail servers' administrators, and that's who you should be asking. They are the only ones who know exactly how their mail servers are configured, and how they work. Unless it's evident from the text of the error message, and it's not, anyone else's answer will be nothing but guesswork.
And actually my guess would be that, if anything, the error message seems to suggest that they have simply blacklisted your IP address range, period, for whatever reason. I would interpret "MX-CIDR" as meaning "MX's IP address' (you can Google what "CIDR" means by yourself); i.e.: sending mail server's IP address is explicitly blacklisted from sending them mail.
Now, taking from the referenced domain's web site, I quote:
"Our software automatically manages the delivery process to give your emails the best chance of landing in the inbox."
I would think that the only type of folks who would be concerned about having "the best chance of landing in" someone inbox would be all the typical spamming parasites. I browsed through the referenced website, and I couldn't shake off a slimy feeling I get after typically wandering into a typical spam spewer.
Is this domain being used to send spam?
If so, then you probably know the answer to your question, already.
Certain SPF libraries might reject emails when trying to perform a reverse lookup on the domain that you're sending from.
They usually get this from the MX records attached to the domain and if there's a mismatch it'll fail out with a rejection (more detail here: http://www.zytrax.com/books/dns/ch9/spf.html).
It's usually only a problem if the receiving server is not necessarily configured correctly, or is being super harsh on incoming mail due to an overwhelming amount of spam.
I send transactional emails from my application on my customers behalf. I'd like for the receiver to be able to click reply and have the reply go directly to my customer, without ever touching my server.
If I send an email from me#myapp.com with the reply-to header set to customer#gmail.com will my transactional emails be flagged as spam? Is this even the slightest red flag to spam filters? It's important my emails don't get caught in the spam folder and i've spent quite a bit of time ensuring that my email is sent properly so I just want to be sure before I implement something like that. Or is this a fairly common practice that I shouldn't be concerned at all about?
I was able to find the answer at another location http://www.quora.com/Does-sending-emails-with-a-From-address-with-a-different-domain-from-the-Reply-To-address-hurt-ones-deliverability
Apparently the reply-to header can be set to another domain without any threat to the deliverability of the email. On some of the smaller mail admins this may cause a problem, but technically speaking its perfectly legitimate.
Provided your sender address is properly set and matches your DKIM/SPF DNS records, you're correct that a reply-to should not negatively effect your sender reputation.
But ultimately it's a provider level decision, so YMMV across different inbox providers.
I've been reading up on sending mass email to a user-base, and I'm not feeling comfortable using the PHP mail function. It tends to be too simple, spammy and unreliable.
But that leads me to my question... for a custom application, what should I be using to send email to potentially hundreds of people? ...or is mail okay to use?
I appreciate the help.
I would use a third-party service. There are several of them. They ensure the emails are sent from white-listed IP's and have spent a LOT of money on legal prep for terms, privacy policy, etc to ensure the ISP's play nicely with the incoming mail.
If you're only sending mail to potentially hundreds of people, and not hundreds of thousands of people, PHP's sendmail will handle the load fine. You should be worrying more about the newsletter content and the opt-out easiness than the capability of PHP to send your email. For small campaigns to hundreds of people, check out MyEmma.com as an example of a small-list solution.
What you're probably looking for is an API to offload your email calls to and let a service handle the delivery for you. Sending a large number of email messages from PHP can be tricky as if it's not done quickly enough you run the risk of time-outs, and tracking which have been sent is always troublesome if you want to re-try a big batch.
Not surprisingly there are several companies which offer an email API service to make this sort of thing significantly easier than doing it yourself:
SendGrid (PHP example)
Mandrill (PHP package) from MailChimp
Postmark (PHP libraries)
PostageApp (PHP example)
MailGun (PHP sample)
While I'm a developer for PostageApp, but I encourage you to try out many of these to see what works best for you.
In most cases you need to re-write a small portion of your application to work with the particular API or library used to access the API, and once that's done you can send a very large number of messages with one quick call. The delivery of those messages becomes the responsibility of your provider.
The truth is that the less money you're willing to spend on email sending, the more things you have to do yourself, such as:
White listing your sender IP address (especially if you're on shared hosting this can be a PITA, because the other users can mess things up for you).
Setting up SPF and DKIM to add trust for mail hosts (Hotmail, GMail, etc.)
Checking bounce emails
Handling ISP complaints
This is also amongst the things that third party providers charge you for; if you don't wish to bother with any of the above, feel free to use providers such as Mailchimp, Bluehornet, etc. Make sure they provide what you need before you whip your wallet, some might have surprising hidden costs (such as charging you extra for API usage, use of transactional emails, life-cycle emails, etc.)
If you don't mind doing a few of the above (like checking bounce / complaint emails and making some simple DNS changes) you could sign up for Amazon SES; it has a proper API and their email charges are the lowest I've seen so far and recently they have introduced DKIM (signed emails) support. You can also configure your sendmail (assuming dedicated hosting) to talk directly to SES, so it's easy to hook up any mail() based solution and run it.
First, thanks for everyone that has helped me with this thus far.
The answer I was looking for is http://mandrillapp.com/
This is the service behind MailChimp and it rules in every way!