Why your emails land in spam — 12 causes and how to fix each one
The spam folder is not a punishment, it is the output of a score. The receiving provider weighs dozens of signals for every message, and most of that score is decided before you press send. This article works through the causes in order of how often they actually occur.
A message does not land in spam because of one mistake. The receiving provider — Gmail, Outlook, a corporate server — scores every incoming message. Part of that score is identity (is this really from where it claims), part is history (what has this sender sent before, how did people react), and a small part is content.
Most companies only work on the third one. They swap words, avoid “free”, delete emoji. But changing content without fixing the first two does not move the result.
1. SPF, DKIM or DMARC is missing
These three DNS records prove the mail really comes from your domain. Since February 2024 Gmail and Yahoo require them from anyone sending more than 5,000 messages a day, and Microsoft published a matching rule for Outlook. If they are missing, your message is filtered before its content is read.
2. High volume from a new domain, with no warm-up
A newly registered domain has no history with the providers. Three hundred messages on its first day does not look like a person; it looks like a bulk sender, and the domain goes onto the suspicious list on day one.
The common misreading here is that the answer is *sending less*. It is not. The answer is climbing to volume slowly: five on day one, ten on day two, fifteen on day three. That pattern tells the receiving side “regular, reasonable volume comes from here.”
When you need more volume, do not push one mailbox harder — add another. And every new mailbox has to start at its own day one; if it inherits the older one’s volume it burns immediately.
3. Spam complaint rate above the threshold
Google publishes the number plainly: keep the spam complaint rate below 0.3%, ideally under 0.1%. That is three people in a thousand pressing the spam button — a far narrower margin than most senders assume.
What pushes that rate up is usually not malice but helplessness: a recipient who cannot unsubscribe has exactly one button left.
4. No working unsubscribe link
Every commercial message needs one-click unsubscribe — both a visible link and an RFC 8058 compliant `List-Unsubscribe` header. This is not only a compliance matter, it is a technical safeguard: a spam complaint costs many times more than an unsubscribe.
5. The list was bought, or it is old
A significant share of addresses in sold lists are closed, and some are spam traps — addresses belonging to nobody, used by providers to catch senders who compile lists without permission. Hitting one trap damages a domain’s reputation in a single send.
6. High bounce rate
Sending to addresses that do not exist tells the provider your list is unverified. Keeping hard bounces under 2% is the widely accepted target. Checking technical validity before an address enters the list removes most of this.
7. Sender name does not match the address
If mail from `sarah@yourcompany.com` displays “Michael”, that is inconsistent for the reader and a weak signal for filters. Every sending address should send under its own name.
A single-word sender name also reads as personal rather than corporate. “Sarah — Company Name” is both recognisable and less suspicious.
8. The email is too long
A twenty-line cold email hurts twice: it does not get read, and it raises the spam score. The same message in ten or twelve lines does better on both counts.
9. Too many links and images
The goal of a cold email is a reply, not a catalogue. Many links, tracking pixels and large images make a message look like a marketing newsletter, and it gets classified as one. Plain text almost always wins in cold outreach.
10. Irregular sending
Sending nothing for three days and then three days’ worth at once is riskier than steady daily volume. Providers reward regularity — which is why sending every day, inside a defined window and distributed across mailboxes, matters.
11. Unfilled placeholders
When a recipient receives `{{company}}` or `[write the industry here]`, that is the clearest possible evidence of bulk templating — to the filter and to the reader. It usually happens when the template changes and the sending code has not caught up.
12. Shared IP pools
On cheap sending services the IP is shared with hundreds of other customers. Another sender behaving badly lowers your delivery rate and there is nothing you can do about it. Sending from your own domain with your own identity removes that risk entirely.
What to fix first
- 1Complete SPF, DKIM and DMARC. Without these, no other improvement changes the outcome.
- 2Put sending on a warm-up schedule, and make sure every mailbox starts at its own step.
- 3Click the unsubscribe link in a real message and confirm it works.
- 4Clean the list: remove addresses with no email, invalid ones, and anyone who opted out.
- 5Shorten the copy, fix the sender name, cut the links.
Breaking this order is the common mistake. A company that works on copy sees no change for weeks, because copy was never the problem.
Frequently asked
My test email arrives fine but real sends go to spam. Why?
An email you send to yourself is judged differently: you have corresponded with that address before, so the provider knows you. To a real recipient the sender is a complete stranger. A test email only proves the formatting is right; it says nothing about deliverability.
How can I see how many of my emails went to spam?
You cannot see it directly — providers do not report it. You use indirect signals: a sudden drop in reply rate, a rise in bounces, and your domain’s reputation data in Google Postmaster Tools. Open rate is not reliable for this, because Apple Mail Privacy Protection inflates it.
Do words like “free” and “discount” really trigger spam filters?
Not on their own. Word lists carry very little weight in modern filters. The same words from an authenticated domain with a steady sending history land in the inbox. Hunting for trigger words is time spent while the real cause stays unfixed.
I have started landing in spam. How long does recovery take?
Reputation falls fast and rises slowly. After the cause is fixed it usually takes a few weeks, during which you should lower volume and climb again gradually. Ignoring the cause and maintaining volume on the same domain only lengthens the recovery.
You do not have to set this up yourself
Tell us your industry and the city you sell to. We find the companies and their business email addresses, and send from your own domain at a pace that keeps you out of spam.