Seed list

Match your actual audience. If most of your recipients are on one provider, a seed list weighted towards three others tells you about a population you do not sell to. Include the providers that matter to you, and include more than one mailbox per provider, because filtering varies with the individual mailbox's history.

Keep the seed mailboxes boring: ordinary accounts with ordinary settings and no aggressive custom rules. A seed mailbox with a hand-written allowlist reports the inbox every time and is worse than useless.

Change one thing at a time. Same content, same sending mailbox, different day tells you about your reputation trend. Different content, same everything else tells you about the content. Changing both leaves you with a result you cannot attribute.

Keep the history. A single placement test is a snapshot with a wide error bar; a series of them is the only way to see a slow drift, which is how deliverability problems normally arrive.

It cannot tell you about a recipient's corporate gateway, their personal rules, or their organisation's block list — a filtered message at one company is invisible to a seed list. It also cannot be extrapolated into a campaign-wide placement rate, however tempting the arithmetic.

Read it as a directional check on your own configuration and reputation. That is a genuinely useful thing to have, and it is less than the number a placement percentage implies.

SalesShift stores a seed list per workspace, runs placement tests against it and keeps the history. With no seed list configured, a test returns a clear no-seed-list outcome rather than an error or an invented figure.