Delivery stops at server acceptance
Campaign platforms usually count a message as delivered when the recipient server accepts it and no hard bounce returns. That is valuable transport evidence. The accepted message may still be placed in Inbox, Promotions, Spam, quarantine, or a user-created folder.
A 98% delivery rate therefore does not mean 98% of recipients saw the message in their main inbox. It means the mail was not rejected in the way the platform counted as a bounce.
Placement observes controlled folders
An inbox placement test sends the production-like message to controlled seed accounts and records the folder outcome. It is good at revealing broad Gmail, Outlook, Yahoo, or business-inbox differences before a campaign. It also keeps delayed or missing seeds visible.
Seeds are a sample, not the audience. They do not reproduce every subscriber's contacts, engagement history, or corporate rule. Placement adds a missing piece of evidence; it does not replace campaign delivery logs.
Use the metrics together
Start with delivery responses when mail is rejected, deferred, or bouncing. Check authentication and infrastructure. Use placement when servers accept the message but opens, replies, or provider patterns suggest filters are hiding it.
After launch, compare delivery errors, complaints, provider reputation, and audience behavior with the pre-send placement result. The combined view can separate a routing failure from a folder failure and a folder failure from an uninteresting offer.
The decision before launch
Send the final message through the real platform, wait for enough seed matches, and review provider results. Delay or repair when an important provider shows a repeatable Spam cluster, the report remains too incomplete, or a hard authentication failure appears.
Record what you accepted. A test becomes valuable history when the team can compare the same sender and stream after the next DNS, platform, audience, or template change.