Why can identical campaigns land in inbox for some users and spam for others?

You send the same campaign to 10,000 people. Some get it in the inbox. Others hit spam. Nobody changed the content. Nobody changed the sender address. So what happened?

The short answer is that inbox placement isn't decided once per campaign. It's decided once per recipient. Mailbox providers like Gmail, Outlook, and Yahoo Mail run their own models on every single delivery, and those models pull from signals that are unique to each user's account.

Here's what actually drives the split.

Engagement history is the biggest one. If a subscriber has opened your last five campaigns, clicked a link, or even replied once, the mailbox provider sees a pattern. This person wants this sender's mail. Inbox it. If another subscriber has ignored every message for six months and once clicked "report spam," the model draws the opposite conclusion. Same email, different outcome.

Sender reputation plays a role too, but it's layered. Your domain reputation and IP reputation are shared signals that affect everyone on your list. But some recipients are on business email hosted by Outlook or Google Workspace, others are on consumer Gmail, and others use Yahoo or ProtonMail. Each provider weighs signals differently. What Gmail considers a trusted sender, a corporate spam gateway at a Fortune 500 company might quarantine based on completely different content rules.

Authentication failures hit some users and not others. If your authentication setup has gaps, the impact isn't uniform. A recipient whose employer runs strict DMARC enforcement will have your email rejected or flagged even if a consumer inbox on the same domain would let it through. SPF misalignments and DKIM failures create unpredictable patterns like this.

Content filtering works differently per provider. Bayesian filters and machine-learning models are trained on what each provider's users have historically marked as spam or engaged with. A subject line that scores fine for one filter might trigger pattern-matching in another. Certain image-to-text ratios, link structures, or even formatting choices can shift placement for a subset of your list.

Local spam folder training matters more than most senders realize. If enough users at a specific domain have previously marked similar emails as spam, the model learns that signal and applies it to new arrivals. This is why a campaign can perform well at Gmail but get buried at Outlook, even when your overall reputation looks clean.

What can you actually do about it?

• Clean up your authentication stack (SPF, DKIM, DMARC). Gaps here create the most unpredictable placement differences across providers.
• Segment by engagement level. Sending to people who haven't opened in 12 months drags down your reputation across the board, not just for those recipients.
• Watch your placement metrics at the domain level, not just overall open rates. If Gmail is showing 40% opens but your Outlook subscribers are at 8%, that's a signal something specific is off for that provider.
• Don't ignore the users who move your email out of spam or mark it as not spam. Those actions feed positive signals back into the model and help future deliveries to similar users.

There's no fix that makes placement identical for everyone. But narrowing the gap starts with knowing which provider is causing the split. If you want to check your authentication setup quickly, our free Email Header Analyzer can show you what each provider is actually seeing when your message arrives.