Monitoring deliverability weekly: the early-warning numbers

Deliverability fails quietly. Your sending tool keeps reporting "delivered" while a growing share of messages lands in spam, and by the time replies dry up the damage is weeks old. A weekly check of five numbers — bounce rate, spam complaint rate, positive reply rate, blacklist status and domain reputation — catches the slide while it is still cheap to reverse.

Why monitor at all?

Because nobody tells you when things go wrong. "Delivered" only means the receiving server accepted the message; an email filed straight into spam still counts as delivered. Mailbox providers do not send you a warning when your reputation slips — they just route more of your mail away from the inbox, gradually. Monitoring is the unglamorous engineering that makes everything else in Cold Email Deliverability: The Practical Guide hold up over months rather than weeks. It assumes your foundations are already in place — if they are not, start with the sending-infrastructure checklist before campaign one and come back.

Which numbers predict trouble?

Five, in order of how early they warn you.

Bounces and complaints are the leading indicators — they move first. Reply rate is the lagging one: by the time it falls, the causes have usually been compounding for a fortnight.

How do you run the weekly check?

The mechanism, as we run it — roughly fifteen minutes per sending domain, same morning every week.

1. Pull bounce rate per campaign and per mailbox. When any figure is above 2%, then pause that campaign and re-verify the list before another send goes out — bounces are a data problem, not a copy problem.
2. Check spam complaints in Google Postmaster Tools and your sending platform. When complaints creep towards 0.1%, then tighten targeting and cut volume; you are emailing people who do not recognise themselves in the message.
3. Check positive replies. When the rate sits below 3% for two weeks running, then treat the campaign as broken and fix the list or the copy — do not simply send more.
4. Run the domain and sending IPs through the major blacklist checkers. When you are listed, then stop sending from that domain and follow a proper delisting process rather than mailing through it.
5. Note domain reputation in Postmaster Tools. When it steps down a band, then halve volume and let engagement recover before scaling again.

Log the five numbers in one row per week. The trend line matters more than any single reading.

What if the problem is not your domain?

Sometimes the red number is inherited. If you send through a platform's shared infrastructure, another tenant's bad behaviour can drag down an IP you happen to share — your authentication is clean, your list is verified, and you are still listed. That is a routing question, not a hygiene one, and it is worth understanding shared vs dedicated IPs for cold email before you blame your own campaign. Volume discipline still applies either way: hold to 25–40 cold emails per day per inbox and scale by adding mailboxes, not by pushing one harder.

Why weekly rather than monthly?

Because reputation moves on a delay, in both directions. A bad list loaded today shows up as bounces this week, complaints next week and inbox placement loss the week after. On a monthly cadence you can spend three of those four weeks sending into spam — at 25–40 emails a day per inbox across several mailboxes, that is hundreds of prospects reached by nobody. Weekly monitoring shortens the loop between cause and correction to days.

One caution: a clean dashboard is not the same as a working pipeline. Monitoring keeps your mail landing; it does not answer it. Most firms stop at two follow-up touches while deals typically need five or more — the follow-up cliff is where well-delivered campaigns still go to die.

---

Next step: the Growth System Audit — £450, seven days, credited against any build — maps where your growth system leaks and what to build first.