Most people find out they have a complaint rate problem the wrong way, because open rates start slipping for no obvious reason, or a deliverability incident lands on someone's desk out of nowhere, or someone finally logs into Google Postmaster Tools after months away and finds a line on a graph that should not be anywhere near that high. By the time it gets noticed, the complaint rate has usually been climbing quietly for a while, and whatever caused it has often moved on to cause something else as well.
What a spam complaint actually is, and what it is not
A spam complaint is a specific action, not a vague outcome, and it is worth being precise about the difference because the two get confused constantly. A complaint happens when someone opens an email and actively marks it as spam or junk, which is a deliberate signal sent back to the inbox provider that this particular sender should not be trusted. An email landing in the spam folder in the first place is a different thing entirely, since that is the provider's own filtering decision rather than a recipient's reaction, and while the two are related, treating them as the same problem leads people to look in the wrong place when something goes wrong. The complaint is the recipient telling the provider something. The folder placement is the provider acting on what it already believes.
When that complaint is logged, it does not sit quietly waiting to be reviewed by a human somewhere. Inbox providers feed it almost immediately into the reputation signals that decide where your next email goes, which means a rising complaint rate is not a lagging indicator you can deal with at your leisure. It is closer to a live alarm.
You need to look at three providers separately, not one
This is where a lot of teams go wrong without realising it, because most ESPs report what looks like a complaint rate, and that number quietly becomes the number everyone trusts. The trouble is that Gmail does not feed its own spam complaint data back to most ESPs, so whatever figure your platform is showing you is, at best, an approximation built from other signals, and at worst something close to a guess. To see what Gmail actually thinks of you, you have to go directly to Google Postmaster Tools, which shows your real spam rate for Gmail recipients, broken down by day, alongside authentication results and reputation bands that your ESP simply cannot see.
Microsoft works differently again. If you are sending from a dedicated IP, Microsoft SNDS gives you complaint and trap-hit data for Outlook and Microsoft 365 recipients, but it reports at the IP level rather than the domain level, and it requires its own separate sign-up that almost nobody bothers with. Yahoo has its own version too, Yahoo Sender Hub, which is smaller in volume for most UK senders but still worth checking if a meaningful slice of your list sits on Yahoo or AOL addresses.
The upshot is that there is no single dashboard that tells you the whole truth. A business can be improving steadily on Gmail while Outlook placement quietly falls apart in the background, and have no way of knowing it from the ESP's own reporting, simply because nobody went and looked at SNDS separately. If you have a dedicated IP, checking all three providers individually is not a nice-to-have. It is the only way to actually know what is happening.
On thresholds: Google's own published guidance puts the danger line at a spam rate above 0.1%, with anything above 0.3% considered a serious problem that will visibly affect inbox placement. Those figures apply to every sender without exception, UK businesses included, and they do not move just because your list is opted in or your subject lines are good.
It is not always about the content of the email
The instinct when complaints rise is to look at the email itself, and ask whether the subject line was too aggressive, or the offer felt pushy, or the send went out on the wrong day. Sometimes there is something to that, but in most cases the email is not actually the problem, and chasing copy fixes for a list quality issue wastes time without moving the number.
People who genuinely want to hear from you rarely mark you as spam, even when the email in front of them is fairly average. The people who reach for the spam button tend to be the ones who lost interest a long time ago, perhaps signed up for one specific reason well over a year ago and have shown nothing since, and for them hitting report spam is simply the fastest way to make something stop arriving. It has very little to do with the specific send that triggered the click.
There is a less obvious cause that catches a surprising number of teams out, particularly anywhere with automated or triggered sends running in the background. A product notification system, an app's lifecycle messaging, or a platform's own triggered flow can each be entirely sensible in isolation, sending something useful and relevant on its own terms. The problem shows up when a recipient moves through a sequence of actions inside the product that the team never modelled properly, and ends up receiving far more emails in a short space of time than anyone designing the system intended. At that point the complaint is not a reaction to any single email's content. It is a reaction to volume, arriving faster than the recipient expected or wanted, regardless of how well each individual message was written.
A more objective way to look at your own complaints
Most teams, understandably, start from the assumption that their own emails are reasonable, which makes it hard to look at rising complaints objectively. A few honest questions tend to surface the real cause faster than guessing does.
- Is the email actually written for the recipient, focused on what is in it for them, or does it read as a list of things the business wants to say about itself?
- A preference centre matters more than most teams assume, since letting people choose between, say, product updates and marketing offers avoids forcing an all-or-nothing decision between everything and nothing.
- The unsubscribe link should be genuinely easy to find, ideally visible and clickable within two or three seconds of opening the email, rather than buried in small grey text at the very bottom.
- Was the original sign-up clear about what the recipient was agreeing to receive, and does the email itself remind them briefly why they are hearing from you, rather than leaving them to wonder?
- Sending frequency tends to creep upward gradually without anyone deciding that on purpose, so it is worth asking whether the current cadence is something a reasonable person would actually expect and welcome.
That unsubscribe point is worth dwelling on for a moment, because it changes the entire calculation for the recipient. Unsubscribing is a polite way of saying no thank you, and it carries no reputation cost to the sender at all. A spam complaint is the opposite: it is harmful to deliverability and it is what happens by default when someone cannot find an easy way to say no. If a recipient cannot locate the unsubscribe link within a few seconds, marking the email as spam becomes the path of least resistance, and that single design choice can be the difference between a clean exit and a damaging signal.
Why the dormant segment makes everything else worse
There is something inbox providers do that catches a lot of teams out, because they do not just score individual emails in isolation (although this balance is changing more recently). They will also score the sender as a whole (as a factor), based on the behaviour of the cohort being sent to. A large dormant segment sitting on your list is not neutral just because it never bounces or complains on any one particular send. It drags down placement for the people who do want to hear from you, every single time you send to everyone at once, simply by being included in the same calculation. Sending to your dead weight hurts delivery to your best subscribers, which tends to be the bit people find hardest to believe until they see it in their own data.
What to do once you understand the signal
A rising complaint rate on its own does not tell you very much, and it should always be read alongside engagement trends, suppression logic, and how the list has actually been managed over time rather than how it was supposed to be managed on paper. Worth checking whenever the number moves the wrong way: which segments are driving it, whether those segments have shown any engagement in the last few months, whether a recent list import or a held-back send happened around the same time, and whether the pattern shows up consistently across Google, Microsoft and Yahoo or only on one of them. A spike isolated to Gmail points somewhere quite different from one appearing everywhere at once.
None of this gets solved by sending less email, and it does not get solved by sending more either. Both are guesses dressed up as strategy. The actual answer comes from understanding your numbers across all three providers and knowing what they are connected to in this specific list and this specific sending history, rather than applying a general rule that happens to apply to everyone.
If you do not know your numbers yet
A surprising number of teams reading this will not actually know their current complaint rate across Google, Microsoft and Yahoo. They also are unaware if their sending set up is Google or Microsoft compliant. If that sounds like you, the place to start is our free health check, which takes about a minute, needs no sign-up, and gives you an immediate read on whether something is structurally wrong with your sending setup.
If the result raises a question or two, or if you already know your numbers and they are not where you would like them to be, book a free inbox check-up call. Fifteen minutes, no obligation, and a proper conversation about what the numbers are actually telling you rather than another guess.
Where this goes next
Monitoring complaint rate is the starting point rather than the fix, and once the data shows a problem, the real answer usually lives in examining engagement cohorts, suppression logic, and how long contacts have been sitting on the list well past the point they should have been removed. Sender reputation sits underneath all of this, and it is worth understanding properly before drawing conclusions from any single number in isolation.
If you want the fuller picture of how a dormant segment does the damage described above, that is exactly what the next post in this series covers.
If your complaint data has already raised real questions, a Deliverability Review and Fix looks at the complete picture, complaint signals, engagement cohorts and suppression logic together, and tells you what is actually driving the problem rather than what the symptom looks like from the outside.
