Cold Email Deliverability Monitoring: Placement Tests Lead, Reply Rates Lag

Two Indicators, Two Jobs

Most cold emailers monitor one number: the reply rate. It is the right number to judge by. It is the wrong number to steer by, because it only moves after the damage is done. A good monitoring setup has both kinds of signal.

Why Waiting for the Reply Rate Burns Domains

Take a plain 2-email drip across 2 days. The first email goes out, the second follows a day later, and replies to both take time to arrive. You cannot tell a quiet campaign from a filtered one until at least 48 hours have passed. That is the floor, on the shortest sequence anyone runs.

Now put a fleet behind it. 100 inboxes at 15 cold emails a day send 3,000 cold emails in those 48 hours. If the domain has gone bad, or the copy has been fingerprinted, every one of them goes to spam, every one teaches the provider that recipients ignore your mail, and the reply rate only tells you once all of it has already happened. Waiting for replies to fall is how domains get burned in two weeks instead of lasting a year.

What a Placement Test Catches Before Replies Do

A bad domain reputation

Google judges cold email largely at the sending domain: every inbox on a domain shares its DKIM key, SPF, DMARC and From domain, which is why one quiet inbox on a healthy domain is a mailbox problem and a whole domain going quiet is a domain problem (one inbox not getting replies, and what we measured in the Google sender reputation study). So a placement test from one inbox reads the whole domain. When it lands in spam on Google, the domain's standing has slipped, and you know before the next 48 hours of sending make it worse.

Fingerprinted copy

Our read of the pattern, which Google and Microsoft do not publish: filters fingerprint a message sent again and again, and once enough recipients ignore or report it, that message gets filtered wherever it comes from. Your DNS is clean, your domain is fine, and the copy is what is burning. A placement test sent with the real campaign copy, spintax rendered, shows it in the spam folder while a test with fresh copy from the same inbox lands. That is the cheapest diagnosis in cold email: swap the copy, not the domain. The Spintax Tester and Spam Checker help you write the next version.

Leveraged Placement Tests: How to Make Them a Leading Indicator

A single placement test is noise. Seed panels differ, filtering is per mailbox, and one email can land in spam by chance (why placement tests disagree). A test becomes a leading indicator when you take the noise out:

• •Send from the real inbox, with the real copy. A pristine template from your best inbox tells you about a campaign you are not running.
• •Best of three. Send it at least three times; the majority decides.
• •Test on your prospects' provider. Google-hosted prospects get a Google test, Microsoft-hosted ones an Outlook test. Our free test at burnedinbox.com runs across several providers at once, or use your own seed accounts.
• •Test on a cadence, not on a hunch. Before every launch and every new variable, weekly on active domains, and after any change: new copy, a new list, a volume step.
• •Compare with yourself. A domain that has tested in the inbox for weeks and now lands in spam is a signal, whatever the absolute number was.

When the Leading Indicator Fires

If most of three tests land in spam: cold sending on that inbox to zero, warmup only for a week, test again. If it lands, ramp back at 5, then 10, then 15 cold emails a day, each step only while replies hold. If it does not, try fresh copy first, then recreate the inboxes on the same domain before you retire the domain. The whole protocol, and how we ran it on 300 inboxes for a year without deleting a domain, is in stop burning cold email inboxes.

Where the Reply Rate Still Gets the Last Word

The test leads; it does not rule. We have seen an inbox test at 100% inbox and not land with real prospects, and an inbox test at 20% and still pull a good positive reply rate. A test that says inbox while replies fall points at the list, the offer or the copy being weak rather than filtered. A test that says spam while replies hold is worth a recheck before you pull anything. When the test and the reply rate agree, act at once.

The Monitoring Stack

• •Underneath, daily: blacklists, SPF, DKIM, DMARC and domain expiry on every domain. The Deliverability Monitor does this free, on inboxes wherever you bought them (closed beta, request access), and you can check any domain now with the Deliverability Checker.
• •Leading: leveraged placement tests on the cadence above.
• •Lagging: reply rate per inbox and per domain against its own last four weeks, and bounces per domain.