Inbox Placement vs. Deliverability: Why "98% Delivered" Tells You Nothing

The most common sentence in deliverability support tickets: "My dashboard says 98% delivered, so why is nobody opening?" Because "delivered" and "in the inbox" are different events, your ESP can only see one of them, and the gap between them is exactly where your revenue is leaking.

Two different events

Delivered means the receiving server accepted the message — it didn't bounce. That's all. Your ESP knows this because the receiving server said "250 OK." A message can be delivered and then filed straight into the spam folder, the Promotions tab, or quietly dropped afterward. It still counts as delivered.

Inbox placement means the message actually landed where a human will see it — the inbox, ideally Primary. No receiving server reports this back. Gmail does not send your ESP a note saying "by the way, that one went to spam."

So when a dashboard shows 98% delivered, the honest reading is: 2% bounced, and for the other 98%, nobody outside Google knows where they landed.

Why nobody can sell you the real number

Folder placement is decided per recipient, using signals nobody outside the ISP can see: that user's personal engagement history with you, their behavior toward similar senders, real-time content classification. There is no API for "what percentage of my mail reached the inbox." Anyone who claims otherwise is estimating — which brings us to seed lists.

Seed lists: useful, but not a measurement

Seed testing sends your campaign to a panel of test inboxes across ISPs and reports where it landed. I use it myself. But understand what it is and isn't:

It's a small, weird sample. A few hundred addresses standing in for your whole list. Seed accounts have no personal engagement history with you — and Gmail personalizes placement per user. Your biggest fan and a seed inbox can receive the same email in different folders.
Seeds don't behave like people. Nobody opens, clicks, replies, or rescues mail from spam on a seed account. The engagement signals that drive modern filtering are absent by design.
It's directional, not decimal. "0 of 40 Gmail seeds hit the inbox" is a real alarm bell. "Inbox placement: 94.2%" is false precision — that number does not exist with that accuracy.

Use seed tests to catch catastrophic changes (a new template that instantly spam-folders everywhere) and to compare A/B variants. Do not report them to your CEO as your inbox rate.

What to watch instead

You can't measure inbox placement directly, but you can triangulate it with signals that are honest about their limits:

Google Postmaster domain reputation + spam rate. This is Gmail grading you in its own words — the single closest thing to ground truth that exists. If reputation is High, your placement is almost certainly fine. If it steps down, placement is degrading no matter what your ESP dashboard says.
Open rate trend, by ISP, with the MPP caveat. Apple Mail Privacy Protection inflates opens, so read Gmail's trend line rather than absolute values. A sustained Gmail open-rate decline is a placement decline — that's usually what it means.
Your own long-lived test accounts. Two or three Gmail accounts you've held for months, that receive every campaign and behave (occasionally opening, never reporting spam). Longitudinal, same accounts, watched over time. Crude, but honest — and unlike seeds, these accounts have history with you.
Complaint rate. Postmaster spam rate plus your ESP's complaint metric. People can't complain about mail they never saw; a falling complaint rate during a volume increase is not always good news — sometimes it means mail stopped reaching the inbox at all.

Why this distinction matters when things break

Deliverability problems and placement problems have different fixes. A bounce spike means list quality or blocklists — fix your acquisition and hygiene. A placement slide with clean delivery means reputation, engagement, or content — a completely different playbook (start with the diagnostic workflow). Treat them as one metric and you'll fix the wrong thing while the right one keeps bleeding.

The uncomfortable summary: "delivered" is a technicality, seed rates are a sample, and the best instruments you have are Postmaster's grade plus your own trend lines. Anyone selling you more certainty than that is selling theater.