Your email platform is reporting 98% delivered, and on the surface that looks fine. The number is green, nothing is on fire, and the sends are going out as planned. But here is the thing: a chunk of those "delivered" emails might be sitting in spam folders right now, completely unread, and your platform has no idea.
Another batch landed in Gmail's Promotions tab somewhere between a flash sale and a subscription renewal notice, competing for attention they will probably never get. Not because your subject line was weak or your send time was off, but because your emails never reached the inbox in the first place.
That gap between what your ESP reports and what is actually happening is email deliverability, and for most growing businesses it is leaking revenue quietly every single day.
"Delivered" does not mean landed in the inbox
When your ESP tells you an email was delivered, it is reporting something very specific: the receiving mail server accepted the message. That is the full extent of it. Think of it like the postman dropping a parcel in the communal hallway of a block of flats rather than placing it directly in your hands. Technically, delivery happened. Whether it reached the right person is a completely separate question.
What your platform is not reporting is whether that email landed in the inbox, the spam folder, the promotions tab, or somewhere else your recipient will never look. All of those outcomes look identical in your delivery report. Your open rate tells you something, but even that has been made considerably noisier by Apple Mail Privacy Protection, which we will come to shortly.
The second thing your ESP is not telling you is your true spam complaint rate. When someone hits "this is spam" in Gmail, Google does not pass that signal back to your sending platform. They keep it and use it to inform filtering decisions for every subsequent email you send from that domain. Unless you are pulling data directly from Google Postmaster Tools, you are not seeing it, which means you can be accumulating a complaint problem without any visible indicator until it is already doing real damage.
Delivery is binary: accepted or rejected. Deliverability is a spectrum, and most businesses are sitting somewhere in the middle of it without knowing.
Shared IPs and the shared laundry problem
When you sign up for a standard sending plan with a lot of ESPs, your emails go out through a shared IP pool, a group of IP addresses used by thousands of senders simultaneously. Your reputation is partly the reputation of everyone else in that pool, which is a more uncomfortable thought the more you consider it.
Picture a communal laundry facility where you and fourteen other households all share the same machines. Your clothes go in for a white wash, but somewhere in the mix somebody throws in a red sock, and by the time the cycle finishes, everyone's whites have come out pink. You did nothing wrong and your laundry was fine, but you were sharing the machine with someone who was not as careful.
The same dynamic plays out on a shared IP. When a handful of senders in your pool start generating high complaint rates, that behaviour affects the reputation of the whole pool. Inbox providers see the IP address before they see your domain, so if that IP carries a history of poor sending, your emails start the journey at a disadvantage through no fault of your own.
Dedicated IPs solve this by giving you your own sending infrastructure, meaning your reputation is entirely yours to build or damage. The catch is that a brand new dedicated IP carries no reputation at all, and no reputation is almost as bad a starting point as a poor one. You have to build it deliberately through a warm-up process, starting with small volumes and gradually increasing as inbox providers learn to trust you. Generally, dedicated IPs make sense above roughly 100,000 emails per month, and only if you have the discipline to maintain consistent sending practices alongside them.
What actually determines whether your email reaches the inbox?
Deliverability is the product of several factors working together, and the important thing to understand is that they interact with each other. Fixing one without looking at the others is why so many deliverability problems come back after being "fixed".
Data quality
Hard bounces, spam traps, and addresses that have not engaged in years are not neutral noise in your list. They actively damage your reputation with every send. Your list is either an asset or a liability depending on which segment you are looking at, and most growing businesses have both living in the same database without a clear plan for separating them.
Consent
This is both a UK GDPR consideration and a deliverability one, and the two are more connected than most businesses appreciate. Contacts who genuinely opted in are more likely to open, more likely to click, and significantly less likely to hit the spam button. Contacts added without a clear, specific consent moment do the opposite. Inbox providers cannot read your privacy policy, but they can read your complaint rates, and those complaint rates are a direct proxy for how the list was built.
Authentication and infrastructure
SPF, DKIM, and DMARC are the technical foundation of inbox trust. Without them properly configured, inbox providers have no reliable way to verify that your email is genuinely from you, and they will treat it accordingly. BIMI, which displays your brand logo directly in the inbox, sits on top of a fully enforced DMARC policy and adds a visible layer of trust for both the provider and the recipient.
Beyond authentication, the architecture of how you send matters considerably. Many businesses, particularly in SaaS and e-commerce, are sending transactional emails (password resets, order confirmations, account notifications) from the same domain and IP as their marketing campaigns. When a marketing campaign generates complaints, those complaints bleed into the reputation of your transactional stream. A promotional email that drives complaints on a Tuesday can suppress the activation email going out on Wednesday, which is a direct and entirely avoidable commercial risk.
Sending behaviour
How you send over time matters as much as what you send. Sudden volume spikes, sending to lists that have not been mailed in months, blasting your entire database for a one-off promotion: these create patterns that inbox providers read as suspicious. Consistent, predictable sending behaviour to engaged audiences is what builds reputation gradually and keeps it stable.
Relevancy is now the whole game, and opens are not the signal you think they are
This is the area where a lot of senders are working with outdated assumptions, and it is worth spending some time on.
The standard way of measuring email engagement has always been built around opens and clicks. Someone who opened three of your last ten emails is "engaged", and someone who has not opened in six months is "dormant". That model made reasonable sense for a long time, but it has some significant cracks in it now.
The biggest crack is Apple Mail Privacy Protection, introduced in 2021. What it does is route email through a proxy server that pre-loads tracking pixels before the recipient ever actually opens the email, which means you can get an "open" recorded in your platform even if the person never read the message. For senders with a high proportion of Apple device users, and across most UK consumer audiences that is a substantial segment, your open rate is now a noisier signal than it has ever been.
There is an important nuance here, though. Opens are not worthless. An email sitting in the spam folder rarely generates an open, even a false Apple-triggered one, because prefetching does not typically apply to mail that has been filtered before delivery. So a complete absence of opens across many sends is still a meaningful negative signal. What you cannot do is treat an open as proof that a real person engaged with your content, because that assumption no longer reliably holds.
The more significant shift is in how inbox providers, particularly Gmail, are now making filtering decisions. Google has moved a long way from rule-based filtering towards machine learning models that assess relevancy at a per-recipient level. The question is no longer just "is this sender generally trusted?" It has become "does this specific person want to hear from this sender right now?" And the signals feeding that model go well beyond opens and clicks.
What this means for how you build your sending programme is significant. If your segmentation and trigger logic are still built primarily around open history, you are working with a signal that Apple MPP has degraded, and which was only ever a partial picture of genuine intent anyway. The more useful approach is to build around first-party signals from your own data: when did this person last log in to the product? What did they do while they were there? Have they visited your pricing page recently? Did they start a trial and not complete it? Did they abandon a basket? These behavioural signals reflect actual intent in a way that an email open simply cannot.
The email that arrives when someone has just revisited your pricing page after a three-week gap is not the same as the email that arrives because their last recorded open was 85 days ago. Both might be labelled a re-engagement email. One of them genuinely is. The other is good timing based on real behaviour, which is a fundamentally different and more effective thing.
Your dormant subscribers are not just dead weight, they are actively causing harm
This is probably the most misunderstood dynamic in email deliverability, and one I see consistently with clients who are genuinely surprised when I explain it.
The conventional thinking is that dormant subscribers are an inefficiency: they cost money to send to and never convert, so the business runs a re-engagement campaign once a year in the hope of waking them up. The problem is that dormant subscribers are not just inefficient. They actively drag down inbox placement for your engaged segment. Inbox providers assess sender reputation partly through engagement signals across your entire sending cohort. When a significant portion of your list never opens, never clicks, and occasionally marks you as spam, that signal is applied at the sender level and affects how the provider treats every email you send, including to the people who genuinely want to hear from you.
Sending to your dead weight hurts delivery to your best subscribers. That is the core commercial point, and most senders have never been told it clearly.
What a sunset policy actually is
A sunset policy is a defined rule for when you stop sending to unengaged contacts. Every brand's version looks slightly different, because what counts as "unengaged" depends on your send frequency, your sector, and how your specific audience behaves. A business emailing weekly will hit a meaningful engagement gap faster than one that mails monthly. The principle is the same regardless: at some point, continuing to send to someone who has given you no signal of interest is doing more harm than good.
Where re-engagement fits in, and why most brands get the order wrong
Re-engagement is often described as the first step in a sunset policy. It is not, and the fact that most brands treat it that way is why so many re-engagement campaigns fail to move the needle.
By the time a subscriber has not opened or clicked anything in twelve months, a "we miss you" email is not going to win them back. You are just adding another low-engagement signal to the pile. They are gone, and the attempts to recover them are making the problem worse for everyone else on your list.
Re-engagement should happen earlier, at the point where you start to see a drop in engagement rather than its complete absence. Catching subscribers while they are still occasionally active, still occasionally visible in your behavioural data, is when re-engagement has a realistic chance of working. And those triggers should be built on the intent signals discussed above, not just on open history. A subscriber who has been quiet in your emails for 60 days but visited your site three times in the last fortnight is not a sunset candidate. They are a warm prospect who deserves a well-timed piece of relevant content.
The right sequence is re-engagement triggered by early signs of declining engagement, then a clean sunset for anyone who does not respond. Not a mass campaign to your entire unengaged list once a year.
The tools that give you some visibility, though not the full picture
There is no single tool that shows you definitively what your deliverability looks like across all inbox providers. What you can do is use several in combination to build a picture, understanding that each one covers a different part of it.
Google Postmaster Tools is the most important starting point for consumer email. It surfaces domain and IP reputation scores for Gmail, the complaint rate Google has recorded (which is different from what your ESP shows), and authentication pass rates. It is free, and if you are not already using it, you should be.
Microsoft SNDS (Smart Network Data Services) is the equivalent for Microsoft properties, covering Outlook, Hotmail, and Live. This matters particularly in the UK, where Microsoft 365 dominates business email environments. Many senders monitor Gmail delivery closely and have no visibility at all into what is happening in Outlook, which is a significant and often expensive blind spot.
Yahoo Sender Hub covers reputation data for Yahoo Mail. It represents a smaller audience than Gmail in most UK contexts but is still worth monitoring at any meaningful sending volume.
Sender Score, provided by Validity, gives a 0-to-100 reputation score for your sending IPs. It is useful as a quick health check and a directional indicator, but it is one data point rather than a definitive verdict on your deliverability.
Tools like GlockApps and Mail Monitor let you send a test email and see where it actually lands across a panel of real inbox providers. Unlike the reputation monitoring tools above, this gives you direct evidence of placement rather than an inferred signal. It is the closest thing available to seeing your email through your recipients' eyes.
The honest caveat: used together, these tools help you build a narrative. You are looking for patterns, anomalies, and signals that warrant investigation before they show up in your revenue numbers. None of them gives you the complete picture on its own.
The part that is hardest to see
Everything above is the surface layer. The problems that cause the most commercial damage and take the longest to diagnose tend to live underneath: in the list acquisition decisions made a year ago, in the trigger logic that has been sending content to people based on signals that no longer mean what they once did, in the infrastructure that was appropriate at one scale and has quietly become a liability at another.
By the time those problems surface in the numbers, they have usually been building for months. The reason they were invisible for so long is that the standard checks showed nothing obviously wrong.
If you want to know what is actually happening with your email programme, the place to start is a proper diagnostic. Not an automated tool report, not a checklist, but a structured investigation that looks at every layer, explains what it finds in plain English, and tells you what it is costing you.
That is what The Digistrat Diagnostic does → Deliverability Audit
Or if you want a quick read on where your setup stands before committing to anything, the free inbox check-up takes 15 minutes and asks for nothing in return. → Book
