Before deliverability, I practiced medicine. I bring that up because it's the honest answer to why my approach to spam placement looks different from most of what you'll find searching this topic. In medicine, you don't treat a symptom. A cough can come from a dozen different places, and treating the cough instead of the cause wastes time while the real problem keeps running. Deliverability works the same way. "My emails are going to spam" is a symptom. It's not a diagnosis, and it's definitely not a fix.
Almost everything written about this topic skips straight from symptom to checklist: check your SPF, check your DKIM, warm up your domain, clean your list. All of that can be correct advice and still not help you, because none of it tells you which item is actually driving your specific problem. You can execute every step on a generic list and still be in spam, because the list was never built around your situation. It was built to be searchable.
This guide is the method I actually use. It runs as six gates, and I stop at the first one that fires. If you want the specific causes behind Gmail placement problems written out as a list, I've done that separately. Here I want to explain something most guides never touch: how to tell which cause is yours, and why the order you check things in changes how fast you find out.
This is the part that makes the problem hard, and it's also the part nobody explains. A domain with a damaged reputation and a domain with a weak engagement problem produce the exact same symptom: mail lands in spam instead of the inbox. So does a shared IP contamination issue. So does a broken DMARC alignment. From where you're sitting, checking your inbox placement, all four look identical. You cannot tell them apart by staring at the result. You can only tell them apart by looking at the evidence behind each one, which means you need to know what evidence to look for and in what order to look for it.
This is the actual reason guessing at a fix so often fails. Not because the fix was wrong in general, but because it was aimed at a cause that wasn't yours. I've had clients rewrite every subject line in their template, migrate ESPs twice, and re-warm a domain from scratch, all while the actual cause was a single misaligned DMARC record that a ten-minute DNS lookup would have caught on day one.
Each gate needs different evidence and moves on a different timeline, and I stop at the first one that fires. The first three carry most cases. The last three exist because a real diagnosis has to account for the cases the first three miss, including the case where nothing is wrong with your sending at all.
Gate 1 · Technical front-loadAlmost every specific cause you'll read about elsewhere, a bad SPF record, a shared IP problem, a stale list, a spam trap hit, sits inside one of those first three gates. Knowing which gate you are at before you start hunting the specific cause is what narrows the search. Guides that jump straight to "check these 12 things" skip this step, which is why they take longer to actually solve anything.
A placement test separates content from infrastructure by sending known-good mail through your own setup and seeing where it lands. It is the gate I run when the first three come back clean or ambiguous, and it is the one no other diagnosis I've seen bothers with. The design of the test is my own and I don't publish it, but what it tells you is straightforward: whether the problem travels with your message or with the pipe it goes through.
Content comes fifth because it is rarely the sole cause and is almost always the first thing blamed. There are real content inputs, and the ones that matter are traceable to published provider requirements rather than folklore. There is no published image-to-text ratio threshold at Gmail, Yahoo or Microsoft, so any number you have been given for one is somebody's guess. A sender with a damaged reputation lands in spam with immaculate copy.
Last, before declaring a fault at all: if one inbox out of many shows spam and everything else is clean, that is the recipient's own filter learning their habits. A sports newsletter landing in spam for someone whose mailbox is otherwise all business mail is not a sender problem, and prescribing a fix for it wastes a month.
I check things in a specific sequence, and it's not the order most guides present them in. I start with whatever is fastest and cheapest to rule out, and I only move to the slower, more expensive checks once the fast ones are eliminated.
Authentication alignment: minutes. A DNS lookup tells you immediately whether SPF, DKIM, and DMARC are configured and aligned. Reputation: an afternoon. Postmaster Tools gives you a banded score and a spam-rate graph you can read the same day. Engagement and list quality: days to weeks. You need to watch actual recipient behavior across multiple sends before a pattern is reliable, and a single send tells you almost nothing.
Worth saying plainly, because it is the part people get backwards: authentication goes first on cost, not on likelihood. Across 56 domains I monitor in Google Postmaster Tools, 51 pass SPF, DKIM and DMARC on 100% of measured mail on a median day, and 2 fall below 95% on any of the three. Authentication is almost never the thing that is broken. It goes first because ten minutes buys you certainty about an entire gate, and there is no other gate where that trade is available.
Content advice, subject line rules, "spam trigger words" to avoid, gets written about constantly because it's easy to turn into a checklist. It's also, in my experience, one of the least common root causes I actually find. A message with a real reputation or engagement problem underneath it will land in spam no matter how carefully the subject line is worded. Chasing the popular advice instead of the likely cause is the single biggest time sink I see.
Deliverability problems aren't always singular. I've diagnosed cases where a domain had a legitimate authentication issue and was also sitting on a contaminated shared IP pool, and fixing only the authentication side produced a small improvement that then plateaued, because the second cause was still active underneath it. Confirming one cause doesn't mean you've found all of them. It means you've found one of them.
You can do real diagnostic work yourself before deciding whether you need outside help. Start with the two fastest gates:
If both of those come back clean and you're still landing in spam, you are at gate three or beyond: engagement, shared infrastructure, or the control test that tells the message apart from the pipe. Those take longer to isolate and are where most of my actual diagnostic work happens. That's also the point where a generic checklist stops being useful, because the remaining gates require reading patterns across your specific sending history rather than matching against a list.
Because a guide gives you a checklist and you need a diagnosis. Working through a generic list fixes whatever happens to be on it, but your cause might not be on that list, or it might be three items down while you started at the top. Deliverability causes have to be ruled out with evidence, in an order set by what is cheapest to check, rather than worked through top to bottom.
Because the checks differ enormously in cost. Authentication resolves in about ten minutes with a DNS lookup. Reputation takes an afternoon inside Google Postmaster Tools. Engagement and list quality need behaviour watched across several sends before any pattern is reliable. Running the cheap checks first means you never spend three weeks investigating engagement decay when the answer was a DMARC alignment failure you could have found before lunch.
No, and that surprises people. Across 56 domains I monitor in Google Postmaster Tools, 51 pass SPF, DKIM and DMARC on 100% of measured mail on a median day, and 2 fall below 95% on any of the three. Authentication is checked first because it is the fastest thing to rule out, not because it is the most likely cause. Ruling it out cheaply is what makes the rest of the order worth running.
The first two checks, yes. You can confirm authentication alignment and read your Google Postmaster Tools reputation bands in under an hour between them. What is harder to self-diagnose is two causes stacked, for example a reputation dip actually driven by a shared IP neighbour rather than your own sending, and the placement test that separates content from infrastructure. Those are where an outside read of the full evidence set matters more than another checklist.
The long-form version of everything on this page, covering all six gates plus what decides placement in the first place, the reputation numbers from my own monitoring, and the tools I read them with, is the complete diagnostic guide.
Each of the first three gates has its own detailed breakdown, written for the specific evidence and failure patterns inside it:
If you've worked through the fast checks and you're still not sure which gate your problem sits at, that's usually the point where a second set of eyes on the full evidence, not another checklist, is what actually moves things. That's the entire premise behind the diagnostic: rule things out in order, with evidence, before touching a fix.