How Does Gmail Determine an Email Is Spam?

Gmail’s junk mail algorithms are quite effective at filtering out unsolicited and unwanted junk email, automatically delivering it to recipients’ spam folders. Unfortunately, there are times when legitimate emails are marked as spam, even though the subscriber wanted to receive them. Although email services such as Gmail don’t share the exact algorithms for deciding if an email is unwanted by intended recipients, the following are some the major reasons why an email may be marked as spam:

• Content — Typically one of the least common causes for pushing an email to the junk box, certain words or phrases found within messages can cause them to be filtered.
• Links — If a message includes links to websites that are blacklisted, they could result in the message being filtered as spam.
• Source — If other email messages sent from the same Internet Protocol (IP) address that you use appear to be spam, your emails may be categorized that way, too. In other words, it’s guilt by association. This can become a problem for small email senders who are on shared servers.
• Headers — If there are inconsistencies in the “from” and “reply to” addresses and domains, emails will be filtered as spam.
• Engagement — Recipients can deliberately mark certain messages as spam, but engagement metrics, such as open and click-through rates, also can affect the email-filtering process. An increasing number of Internet service providers (ISPs), including Yahoo! and Gmail, are using such engagement metrics to determine if an email should automatically be blocked or filtered as spam. In effect, Google is simply trying to provide a great user experience, just as Google’s search algorithms do. So, Google views engagement metrics as an indicator of whether or not the user is interested in emails from that sender. We have seen tests where lists with high levels of engagement get through Gmail’s junk filters, while lists that without the engagement don’t get through.

Last point make sure you have a SPF record from the ip/domain where it leaves, also A PTR record for the broadband your using as sometimes we do reverse look ups.

Hope this helps

Az