Your ESP dashboard says 98% delivered. Your open rate is 4%. If you've ever stared at that gap and wondered what happened, the answer is almost always the same: Gmail delivered the message, then filed it into a tab your recipient never checks. Primary, Promotions, Social, Updates, Forums, Spam — six boxes, six very different engagement outcomes, one indistinguishable "delivered" status upstream.
ESP delivered ≠ inbox placed. A free Gmail tab placement test sends your message to real seed mailboxes, reads which tab Gmail chose, and returns screenshots plus the SPF/DKIM/DMARC verdict. GlockApps charges $59/mo after 3 free tests. Inbox Check is unlimited and free.
What each Gmail tab actually is
Gmail introduced tabs in 2013 and quietly re-tuned them at least four times since. The six buckets an incoming message can land in:
- Primary — person-to-person and mail Gmail cannot confidently classify as bulk. Highest engagement, highest attention.
- Promotions — marketing, deals, newsletters with marketing DNA. Not spam, but the user actively chooses when to look.
- Social — notifications from social networks, dating, media sharing sites.
- Updates — receipts, confirmations, statements, automated transactional mail.
- Forums — mailing lists and discussion groups.
- Spam — filtered out. Most users never look.
The tabs are not a punishment and not a reward — they are Gmail's attempt to shape attention. A B2B cold email in Promotions is a bug. A receipt in Primary is also a bug (from Gmail's perspective).
Why your ESP can't tell you the tab
SMTP has no concept of tabs. The receiving server issues a 250 2.0.0 OK and the ESP records "delivered". What happens inside the Gmail UI after that — tab routing, clipping, image caching, the "Be careful with this message" banner — is completely invisible to the sender. Webhooks do not fire on tab placement. Nothing in the bounce/complaint pipeline carries it.
The only way to know the tab is to look inside a real Gmail inbox. That means seed accounts.
The signals Gmail uses to classify
From public Postmaster documentation, Workspace engineering talks and pattern analysis across thousands of placement tests, the main classifier inputs are:
List-Id,List-UnsubscribeandPrecedence: bulkheaders — the clearest bulk signal.- From-domain reputation, specifically engagement rate across the recipient population.
- Content DNA: CTA buttons, image-to-text ratio, tracking pixels, branded headers and footers, marketing vocabulary.
- Sending pattern: a domain that sends 50,000 messages per day from an ESP IP pool looks very different from a founder sending 30 a day from Workspace.
- Per-user engagement history — one recipient may train Gmail to put your mail in Primary while another trains it into Promotions.
How a free placement test reveals the tab
A seed-based test works like this:
- You grab a one-shot BCC address from the test tool.
- You send your real campaign and BCC the seed list.
- Each seed mailbox authenticates against Gmail (consumer and Workspace), Outlook, Yahoo, Mail.ru, Yandex, GMX, Orange, ProtonMail, iCloud — 20+ providers in total.
- The tool reads which folder / tab the message arrived in and captures a screenshot.
- Alongside the placement, it parses Authentication-Results and runs SpamAssassin + Rspamd on the raw MIME — so you see SPF, DKIM, DMARC alignment and content score in one view.
Interpreting screenshot-level results
When the result comes back, read it in this order:
- Authentication block. If SPF, DKIM or DMARC failed, stop reading content analysis — fix auth first. Promotions placement is normal with clean auth; Spam placement without passing auth is probably the auth.
- SpamAssassin / Rspamd score. Above 5.0 and you have a content problem. Below 1.0 and you have a reputation or engagement problem, not content.
- Tab distribution. If 8/10 Gmail seeds land in Promotions but 0 in Spam, you're shipping marketing mail and Gmail knows it. That's a product/content question, not a deliverability one.
- Screenshots. Look for the "Be careful" banner, a clipped body, image blocking, or the Workspace admin warning — each tells a separate story.
GlockApps vs Inbox Check at a glance
GlockApps Inbox Insight: 3 free tests, then $59/month for the entry plan. ~60 seed accounts, heavy US/EU skew.
Inbox Check: unlimited free tests, no signup, no credit card. 20+ providers including Gmail, Outlook, Yahoo, Mail.ru, Yandex, GMX, Orange, ProtonMail, iCloud. Real screenshots. Public MCP endpoint for Claude / ChatGPT automation.
4 common fixes for Promotions-exiled senders
- Strip the marketing chrome. Tracking pixel, multi-column layout, 600px hero image, four CTA buttons — each one pushes you toward Promotions. A plain-looking text-forward email ranks higher in Primary.
- Authenticate strictly. DKIM with a 2048-bit key, DMARC at
p=quarantineorp=reject, SPF with-all. Gmail trusts strictly-aligned senders more. - Send from a person, not a brand.
sarah@company.comoutperformshello@company.comin Primary routing. - Ask engaged recipients to drag you to Primary once. The signal is small per-user but compounds across your list.