Email reputation monitoring: what to check, how often, who owns it

The difference between a bad week and a bad quarter is how early you notice. Reputation damage is gradual and visible: bounce rate creeps, one provider's placement slips, complaints tick up on a particular segment. By the time reply rates fall enough for someone to ask why, the cause is three weeks old and the recovery is weeks more. A monitoring routine fixes this, and it is short enough that one person can own it.

Daily, five minutes

• Hard bounce rate on yesterday's sends. Under 1% is healthy. Over 2% means stop and clean the list before sending again.
• Complaint rate where your provider reports it. Google publishes 0.3% as the threshold bulk senders must stay under in Postmaster Tools, and recommends staying below 0.1%.
• New 4xx and 5xx patterns. Not the count — the strings. A new 421 4.7.0 from one provider or a 550 5.7.1 naming a blocklist is the earliest actionable signal you get.
• Volume against plan. A send that is 5x yesterday's, intended or not, is worth knowing about before the provider tells you.

Weekly, thirty minutes

• Per-provider inbox placement from seed mailboxes. The aggregate number is useless; the per-provider split is the whole point. One provider well below the others is a block or a reputation split.
• Gmail spam rate trend in Postmaster Tools, plus the authentication and delivery-error panels.
• Per-mailbox outliers. On a fleet, compare mailboxes against each other. One mailbox diverging is usually a compromised account or a worse list, and it is invisible in a domain-level average.
• Blocklist state for your sending IPs, your sending domains, and the domains in your links — tracking and redirect domains carry their own reputation.
• Unsubscribe and reply handling. Confirm requests are actually being honoured and replies are reaching a human.

Monthly, a couple of hours

• Authentication audit. SPF under ten DNS lookups and no stale include:, DKIM signing with a key you rotate on a schedule, DMARC aggregate reports read rather than filed.
• Sender inventory. Every system allowed to send as your domain, compared against the list from last month. The thing that causes incidents is usually a tool someone connected last year.
• List hygiene. Suppress addresses that have not engaged in your chosen window, and remove role accounts you never should have mailed.
• Stream separation. Confirm transactional, marketing and outbound still live on separate domains or subdomains and nobody has "temporarily" merged them.

Thresholds worth writing down

Write them down because the alternative is a judgement call made by whoever happens to look, and judgement calls drift upward.

What you can no longer monitor

Google Postmaster Tools retired its domain and IP reputation charts, so there is no reputation grade to track as a metric any more. Any runbook, dashboard or client report still built around "domain reputation: High" needs replacing. The substitute is a pair of things you already have in the lists above: Gmail spam rate from Postmaster Tools, and per-provider seed placement everywhere else.

One honest caveat on the placement half. At Gmail, the Promotions tab carries the INBOX label, so any measurement built on reading labels counts a Promotions landing as a clean inbox hit. Treat label-derived placement numbers as an optimistic ceiling, and eyeball the tab on at least one seed before you celebrate.

What not to do

Do not monitor a third-party sender score as your KPI. No mailbox provider reads it, and you can spend months moving a number with no effect on where your mail lands. Do not build alerting you then mute — one noisy alert everybody ignores is worse than none, because it provides the feeling of coverage. Do not wait for a listing before reacting; most reputation damage never produces one. Do not check only aggregates, which is how a single bad mailbox or a single blocked provider hides for a month. And do not let monitoring become the work: the point is to catch three or four things a year early, not to produce a weekly report nobody acts on.

Give it an owner

Monitoring without a named owner decays within two months. One person, five minutes a day, with written thresholds and the authority to stop a send. That authority is the part that matters — a monitor who can see a problem but cannot pause the campaign is just documenting the damage.

MailPilot automates most of the daily and weekly list: it measures real per-provider inbox placement through a seed network, records per-mailbox reputation daily so outliers are visible against the fleet, and checks SPF, DKIM and DMARC on every connected domain.