What is a seed list?

Build one with real accounts at Gmail, Outlook, Yahoo, Apple, plus whatever corporate systems your buyers actually use. Ten accounts spread across providers tell you more than fifty at one.

Keep the seeds out of your reported metrics. A handful of guaranteed opens on a small send skews the rate enough to hide a real problem.

What the test is genuinely good for is change detection. Placement that was fine on Tuesday and is spam on Friday points at something specific: a new link domain, a content change, an authentication break.

Commercial seed panels offer hundreds of accounts and a dashboard. They cost real money, and they suffer the same blind spot as a homemade list, because their accounts have no engagement history with you either.

Here's the routine for a homemade list of twelve accounts. Add them to the campaign as a tagged segment. An hour after the send, log in to each one and record three things in a spreadsheet: the folder, any warning banner, and the authentication results from the message source.

Say this week's sheet reads inbox for eleven accounts and spam for one Yahoo mailbox. A single outlier isn't a finding. Send again next week and see whether it repeats. If the pattern is the same three weeks running, it's worth your time, and a one-off is just noise.

The history is where the value builds up. Six months of weekly rows let you connect a change in placement to a change you made. Perhaps placement at Outlook dropped the week you switched link tracking domains, and recovered when you switched back. You can't get that comparison from one test, however many accounts it uses.

Seed accounts don't open or click, which makes them look exactly like the dead weight your sunset policy is meant to remove. Unless they're exempt, an automated rule that drops 90-day non-openers will delete your whole seed list in the first quarter. You'll only notice when someone asks why the spreadsheet stopped filling in.

The opposite mistake is just as easy to make. If someone on the team opens the seed messages to check rendering, the accounts start building a positive history with you. Their placement improves, and they stop representing a cold recipient.

So choose a purpose for each account and hold to it. Tag seeds as exempt from engagement rules, keep them out of reported metrics, and keep rendering checks on separate accounts. Don't mistake a seed list for a verifier test file, either. A seed list measures placement after you send. To test a verifier you need addresses whose status you already know, and you score how often its verdicts agree.