Cold email going to spam: isolate the failing layer

Define the decision

One test account is a useful symptom, but it is not a representative sample of every recipient. Preserve the original message, full headers, sending mailbox and observation time before making changes that erase your baseline.

Work through the procedure

1. Determine whether the message was rejected, accepted into spam or absent from the receiving account.
2. Check authentication on the received copy, not only DNS scanner results.
3. Compare the affected provider, audience source and campaign version with an unaffected control.
4. Change one plausible cause and repeat a controlled observation before expanding.

Worked example

The following is a synthetic example for this procedure, not a customer result or performance benchmark.

Control: same mailbox, relevant known recipient, no new attachment
Variant: current campaign message
Observation: only the variant lands in spam
Next investigation: message links and construction, while retaining authentication evidence

Read the result

This comparison narrows the next investigation; it does not prove that a single word triggered a filter. Recipient history, provider rules and timing can still differ. Record what remained constant as carefully as what changed.

Check before moving on

1. Exclude warmup exchanges from prospect-placement claims.
2. Preserve negative replies and opt-outs as audience-quality evidence.

Limits and next action

Avoid repeatedly sending to unresponsive people to obtain more data. Pause an affected segment when the cause remains uncertain and use provider diagnostics alongside controlled tests.

Source: Google: email sender guidelines