Email deliverability testing answers one question: does mail from this domain and mailbox reach the inbox, or does it land in spam, Promotions, or nowhere at all? Open rate and click rate do not answer that question directly - they are downstream proxies, and both have gotten noisier over the past few years.
Why open rate stopped being a reliable signal
Apple's Mail Privacy Protection, active since iOS 15, pre-fetches remote images for a large share of Apple Mail users regardless of whether the recipient opens the email. That inflates open rate for any list with meaningful Apple Mail share, and there is no way to separate a real open from a protection-driven pre-fetch inside the metric itself. Gmail's image proxy does something similar at a smaller scale. Treat open rate as a rough trend indicator across your own history, not an absolute number, and never as proof of inbox placement.
The four metrics that actually indicate deliverability
| Metric | What it tells you | How to get it |
|---|---|---|
| Inbox placement rate | Share of sent mail landing in inbox vs. spam vs. Promotions, per provider | Seed-inbox test (send to known addresses, check folder) |
| Authentication pass rate | Whether SPF, DKIM, and DMARC align on outbound mail | DNS lookup + header inspection on a real sent message |
| Spam complaint rate | Recipients actively marking mail as spam | Feedback loop data from the sending platform (e.g. Gmail Postmaster Tools, ESP complaint reports) |
| Sender/domain reputation score | A provider's running assessment of your sending history | Google Postmaster Tools (domain reputation), Microsoft SNDS |
Spam complaint rate deserves the most attention of the four. Google's long-published guidance treats a complaint rate above roughly 0.3% as a threshold that risks bulk-sender consequences - that is the one number in this space backed by a figure Google itself has published, rather than an industry estimate.
How a seed-inbox test actually works
A seed-inbox network is a set of real mailboxes across Gmail, Outlook, Yahoo, and other major providers that you control or a tool controls on your behalf. You send a real email to the full set, wait for delivery, then check each inbox (inbox, spam, or - for Gmail - Promotions and Updates tabs) to see exactly where that message landed. Run this before a volume increase, after any authentication change, and periodically during a sustained campaign - placement can shift week to week even with no changes on your end, because provider-side filtering models update continuously.
Example seed test output gmail-seed-1@... -> Inbox gmail-seed-2@... -> Spam outlook-seed-1@... -> Inbox outlook-seed-2@... -> Inbox yahoo-seed-1@... -> Inbox
A result like this - one of two Gmail seeds in spam - is a real signal worth investigating (new domain reputation, recent complaint, or a content trigger), not noise.
What testing tools can and cannot tell you
A spam-checker tool that scores a single email against common filter heuristics (spammy words, broken HTML, missing authentication) is useful as a pre-send sanity check, but it is not deliverability testing - it does not touch real mailbox providers or reflect your sending domain's actual reputation. Treat it as a linter, not a placement test.
Building a weekly testing habit
Deliverability testing is most useful as a recurring check, not a one-time audit. A reasonable cadence: run a seed-inbox test weekly during steady-state sending, after any change to sending domain or volume, and immediately after any authentication change (a new DKIM key, a DMARC policy tightening from p=none to p=quarantine). Log the results over time rather than treating each test in isolation - a single Gmail seed landing in spam once might be noise, but the same seed landing in spam three weeks running is a trend worth investigating before it spreads to the rest of your Gmail-hosted recipients.
What not to do
Do not rely on a single send to judge deliverability. One blocked message can be a fluke (temporary greylisting, a transient provider-side block); one clean inbox placement does not guarantee the next hundred sends will land the same way. Test repeatedly, and track the trend, not the single data point.
Do not chase open rate improvements as a proxy for fixing deliverability. If placement is the actual problem, a better subject line will not move it - only authentication, reputation, and content changes will.
Where MailPilot fits
MailPilot runs a seed-inbox network across major providers to show real inbox placement, not a proxy for it, alongside SPF/DKIM/DMARC verification against your live DNS record and a daily reputation score per mailbox. It connects over OAuth (Gmail, Outlook) or SMTP/IMAP (SendGrid, Amazon SES, Mailgun, Postmark, custom SMTP) - the testing layer sits on top of whichever of those you are already sending through, rather than replacing your ESP.
Ready to reach the inbox every time?
Automated email warmup across real peer mailboxes. Live inbox placement monitoring. Free 7-day trial - no credit card required.
Start free trial