Why Your Emails Are Landing in Spam Even Though Nothing Changed

"We haven't changed anything" is the most common opening line in a deliverability investigation, and it's almost always true and almost always the wrong place to look. Deliverability rarely collapses because of one obvious, dramatic mistake that someone can point to. It degrades because several independent inputs - authentication, list health, recipient engagement, content patterns - each drift slightly over time or shift under conditions outside the sending program itself, and the combined effect eventually crosses a threshold that a mailbox provider's filters act on. From inside the sending program, nothing looks different. From the mailbox provider's side, enough has changed that the filtering decision changes with it.

Deliverability Is an Outcome, Not a Setting

The first thing worth separating clearly: deliverability is not a dial anyone turns directly. It's the observed result of several underlying factors working together, and no single one of them fully explains a deliverability problem on its own. Treating deliverability and sender reputation as interchangeable - "our reputation must have dropped" - skips past several other levers that are just as capable of tanking inbox placement even with a technically healthy reputation score.

The Inputs That Actually Drive It

Authentication setup

SPF, DKIM, and DMARC records tell a receiving mail server that an email claiming to come from your domain is actually authorised to do so. A misaligned or missing record on any of the three doesn't just risk one email - it can affect deliverability across an entire domain's sending, and it's the kind of thing that can break silently. Adding a new sending tool, changing a DNS provider, or migrating a subdomain can all quietly break alignment without triggering any obvious error in the sending platform itself.

Sender reputation

This is the factor most people jump to first, and it matters - but modern mailbox providers, particularly Gmail and Outlook, increasingly weight recipient-level engagement behaviour as part of reputation rather than relying purely on static domain and IP history. Opens, replies, and manual "not spam" corrections from real recipients feed into reputation continuously, which means reputation isn't just something that was earned once and now sits fixed - it moves with how people are actually behaving toward recent sends.

List hygiene and engagement decay

A list that grows steadily but keeps a large base of long-inactive addresses dilutes the overall engagement rate a mailbox provider sees for that sending domain. Continuing to email recipients who haven't opened anything in a year, even if they never explicitly unsubscribed, drags down the aggregate signal mailbox providers use to judge whether a domain sends wanted mail.

Content and structural spam triggers

Spam filters still weigh content and structural signals - excessive links, certain phrase patterns, a poor text-to-image ratio, broken or suspicious tracking links - alongside reputation and engagement. A single subject line or template change can shift a specific campaign's filtering outcome even when nothing about the sending domain's underlying history has moved at all.

Why "Nothing Changed" Is Rarely True

Recipient behaviour changed even if sending didn't

A segment that's gone quiet over several months, then gets included again in a broader send, brings its low engagement rate back into the aggregate the mailbox provider is measuring - dragging down the metrics for that specific send even though the sending program's configuration is identical to the last successful campaign.

A new integration silently added a sending path

Connecting a new tool that sends transactional or marketing email through the same domain - a helpdesk tool, a new CRM automation, a form notification service - can introduce a sending source that isn't properly included in the domain's SPF record or aligned with DKIM, degrading authentication for the whole domain without anyone touching the original marketing platform's settings at all.

List growth diluted engagement even as volume grew

A list that doubled in size over a year while engagement rate held flat in absolute terms can still represent a falling engagement percentage if growth outpaced how quickly new subscribers actually start opening mail. Total sends looking healthy on a dashboard can mask an engagement ratio that's quietly moved in the wrong direction.

A Concrete Example

Take a company whose weekly newsletter had landed reliably in the inbox for over a year, sent from the same domain, same platform, same general format. Inbox placement started slipping gradually over a few months - not a cliff, just a slow decline that showed up first as a falling open rate before anyone suspected deliverability specifically. Nothing in the sending platform's configuration had changed, and the team's first instinct was to check sender reputation through a monitoring tool, which came back looking unremarkable.

The actual cause turned out to be two unrelated things compounding at once. First, a support tool integrated eight months earlier had been sending automated notification emails from a subdomain of the same root domain, without ever being added to the domain's SPF record - a gap nobody had reason to check, since the support tool's own emails weren't the ones losing deliverability. Second, the newsletter list had grown by re-including a large batch of contacts who'd been dormant for over a year, diluting the aggregate engagement rate mailbox providers were measuring for that specific domain. Neither issue alone would likely have caused a visible problem; together, they pushed enough negative signal into the domain's overall standing that even the previously reliable newsletter started losing inbox placement. Fixing the SPF gap and re-segmenting the dormant contacts out of the main send resolved it within a few sending cycles - not because either fix alone was dramatic, but because both of the actual causes were finally addressed instead of the reputation score everyone had checked first.

A Practical Diagnostic Order

When deliverability drops, checking things in a deliberate order avoids chasing the wrong lever first. Start with authentication - confirm SPF, DKIM, and DMARC are correctly configured and aligned, since a break here can invalidate almost anything else that's investigated afterward. Next, check reputation and blocklist status directly through the sending platform or a dedicated monitoring tool, rather than assuming. Then look at engagement segmentation - is the drop concentrated in a specific segment, or spread evenly across the whole list. Finally, review the specific campaign's content and structure for anything that might trigger filters independent of the broader sending history. Working in that order surfaces the actual cause far faster than guessing at reputation first and stopping there.

Building Deliverability Resilience Long-Term

A few ongoing practices reduce how often this investigation needs to happen at all. Warming up any new sending domain or IP gradually, rather than sending full volume from day one, gives mailbox providers a chance to build a positive history before it's tested at scale. Regular list hygiene - actively sunsetting or re-permissioning subscribers who haven't engaged in a defined window, rather than letting them sit indefinitely - keeps the aggregate engagement rate mailbox providers see closer to the program's actual healthy core. Segmenting sends by engagement level, so highly active and long-dormant recipients aren't blended into the same broad campaign, protects the sender's reputation signal from being dragged down by the quiet portion of the list. And monitoring postmaster tools and feedback loops where available gives an early warning before a slow decline becomes a visible drop, rather than finding out only once inbox placement has already suffered.

None of these individually guarantees deliverability, and that's the point worth sitting with: it was never one setting to protect, but several independent ones that all need to stay healthy at the same time. A program that keeps all of them in reasonable shape rarely faces the "nothing changed" mystery in the first place, because the small drifts that would otherwise compound get caught and corrected long before they add up to a filtering decision.