Sending reputation is the asset. Protect it first.

Connect every mailbox you intend to send from

The pool is built from the mailboxes your organization has connected — the same ones from Settings → Integrations and Settings → Mail. A pool of one is legitimate and often correct; rotation is for teams sending enough volume that a single mailbox would hit its cap.

Read the pool table

The pool view lists every connected mailbox with its cap, ramp position, health and — when it is not sending — the reason. Alongside it are the org totals: how many mailboxes are pooled, how many are actually sendable, how many are paused, how many need reconnecting, today's capacity, what is left of it, the org-wide daily ceiling, and which mailbox is up next.

• ▸The health numbers are the last computed snapshot, not a fresh calculation. The page tells you when they were computed, how old that is, and whether a refresh is running.
• ▸That is deliberate. Computing health inline would couple page latency to write volume, and a page that times out shows you nothing at all — which is worse than slightly older numbers, as long as it says so.
• ▸A mailbox that needs reconnecting is counted separately from a paused one, because the fix is different: one needs a new consent, the other needs a decision.

Set a cap and a warmup ramp per mailbox

Each mailbox carries its own policy: whether it is in the pool, its weight in the rotation, a daily cap, a ramp start date with an initial volume and a daily increment, and a minimum gap between sends. The ramp is the important one for a new mailbox — it starts you low and adds a fixed amount per day instead of letting a cold mailbox send at full cap on day one.

• ▸The org defaults are shown on the same page, so you can see what a mailbox inherits before you override it.
• ▸Pause state is deliberately not editable in this policy form. Pausing and resuming have their own actions so that lifting a pause always records who lifted it.

Turn rotation on — as an explicit decision

Rotation ships OFF. Turning it on changes which address your mail comes from, and doing that to a live domain without being asked is a good way to damage a sending reputation. So it is a switch someone flips once, knowingly, after confirming the mailboxes are configured. The same panel sets the health window in days.

Fix domain authentication before you test anything

The domains view lists every sending domain your mailboxes use, with the last DNS answer for each: SPF, DKIM, DMARC, MX and blocklist status. A domain with no check timestamp has never been checked, and the page says exactly that rather than showing a green tick by default.

• ▸Answers are served from cache while fresh; a refresh re-queries DNS now.
• ▸If the DNS engine is unreachable but a previous answer exists, that answer comes back flagged stale with the error attached. A six-hour-old DMARC record is useful and clearly labelled; an invented one is not.
• ▸DKIM needs the selector to be checkable. Without one it is reported as unknown, not as passing.

Build a seed list you actually control

A seed address is a mailbox that receives placement probes. Add several across the providers your prospects use. The seed list is joined to your connected webmail accounts, which matters for the next step.

Run an inbox-placement test

A placement test sends a probe to every seed and reports where it landed. Two behaviours are worth knowing before you read the result. With no seed list, the test returns a "no seed list" status and an explanation. A placement test with nothing to send to has no result, and printing a made-up inbox percentage would be worse than printing nothing. Placement itself is only claimed for seeds that are mailboxes your organization has connected — reading the destination folder is the only way to know where something landed. Every other seed is recorded as sent, with placement left null.

Watch it while you send

Health is scored over a rolling window from complaint rate, bounce rate and failure rate, with a bonus for engagement, and the page publishes the formula it used — the base score, the window, the minimum sample before a score means anything, each penalty, and the grade boundaries. It also publishes the auto-pause thresholds, so you know in advance what will pull a mailbox out of rotation and what it takes to bring it back.

• ▸A mailbox below the minimum sample size does not get a confident grade. Small numbers are noise.
• ▸When a mailbox is paused, the reason is recorded with it — including when the pause was automatic.