How does Gmail determine placement?

You hit send, and 30 seconds later your email is sitting in the Promotions tab instead of the inbox. Or worse, spam. What just happened?

Gmail doesn't use a single rule or checklist. It runs a machine learning model that weighs dozens of signals at once, and the weighting shifts constantly as the model learns from its 1.8 billion users. That said, there are clear layers to how it works.

Layer 1: Authentication (the floor, not the ceiling)

Before Gmail evaluates anything else, it checks whether your email is actually from you. That means SPF, DKIM, and DMARC need to pass and align. Failing authentication doesn't automatically send you to spam, but it removes a layer of trust that Gmail uses to vouch for your domain. Think of it as the baseline. Without it, you're starting at a disadvantage before the model even sees your content.

Layer 2: Sender reputation (your history with Gmail)

Gmail maintains a reputation score for your sending domain and IP address. This score is built from aggregate signals across all Gmail users, not just your own list. It tracks things like how often your emails are marked as spam, how often they're deleted without being opened, and whether your domain has shown erratic sending patterns. A domain with a long, clean sending history has a reputation buffer. A brand-new domain or one recovering from a deliverability crisis has almost none.

Layer 3: Per-user engagement (this is where Gmail gets personal)

This is the part most senders underestimate. Gmail's model is personalized. Two people on your list can receive the exact same email, and Gmail might send it to the inbox for one and to Promotions for the other, purely based on how that specific person has interacted with your domain before.

Positive signals include opening regularly, clicking, replying, starring, and moving an email from one folder to another. Negative signals include ignoring, deleting without opening, and spam reports. A spam report is by far the most damaging single action a recipient can take. Even a handful of reports on a small list can push your reputation into the red.

Layer 4: Content and behavioral prediction

Gmail also scans the content of your email and compares it against patterns it associates with spam, phishing, or low-quality sending. This doesn't mean your subject line needs to be sanitized. It means things like suspicious link patterns, mismatched sender names, or content that looks like known scam templates will trigger filters. Gmail also predicts whether a given user is likely to want this message, based on what similar users have done with similar messages in the past.

What you can actually control

Authentication is table stakes. Get SPF, DKIM, and DMARC in place and don't touch them once they're working. Beyond that, the most powerful thing you can do is earn genuine engagement. Send to people who actually want to hear from you, keep your list clean by removing people who've stopped responding, and make sure every email you send has a reason to exist. Gmail's model is very good at detecting the difference between an email someone is happy to see and one they're tolerating.

But if you want to see how your authentication is sitting right now, our free SPF checker takes about 30 seconds to run. Or if you're troubleshooting a placement problem and want a second set of eyes, the SOS hotline is free.