Inbox placement: the number delivery rate hides

What inbox placement is

Inbox placement is the percentage of your sent mail that lands in recipients' inboxes — as opposed to the spam folder, a categorized tab, or nowhere at all. It is the metric that describes what your audience actually experiences, and it is the one number most email dashboards cannot show you.

The reason they cannot is structural. When a mailbox provider accepts a message over SMTP, it tells the sending server "accepted" and nothing more. The filtering decision — inbox, spam, Promotions tab, or silent discard — happens after that handshake, inside the provider's infrastructure, and no status code reports it back. Your ESP's "delivered: 99.2%" means 99.2% of messages were accepted for delivery. Every one of them could be sitting in spam and the dashboard would look identical.

Measuring placement therefore requires being on the receiving side: seed mailboxes at each provider, real campaigns sent to them, and a count of where each message landed. That is the method behind every deliverability number on this site, documented in our test methodology.

How inbox placement differs from delivery rate

The two metrics answer different questions and are routinely conflated, including by ESP marketing pages.

Delivery rate answers: of the messages we tried to send, how many did receiving servers accept? It is measured at the SMTP boundary, visible to the sender, and typically high — bounces are the only thing that lowers it.

Inbox placement answers: of the accepted messages, how many did a human have a realistic chance of seeing? It is measured post-acceptance, invisible to the sender without seed data, and varies far more between senders and providers.

A vendor claiming "99% deliverability" is almost always citing acceptance, not placement. The two can diverge without limit: a domain with a damaged reputation can hold a 99% delivery rate and a 40% inbox rate simultaneously. When we publish placement, we break out the states separately — inbox, categorized tabs, spam, missing, bounced — because collapsing them is how the misleading numbers get made.

How inbox placement works

Placement is decided per message, per recipient, at filtering time. The major inputs:

• Sender reputation — the history attached to your domain and IP, built from complaints, engagement, spam-trap hits and volume patterns.
• Authentication. SPF, DKIM and DMARC results, with aligned authentication now a bulk sender requirement at Gmail and Yahoo.
• Engagement history with this specific recipient. Providers personalize: the same campaign can inbox for a subscriber who opens weekly and junk for one who has ignored forty sends.
• Content and structure, a smaller factor than folklore suggests, but real at the margins.

Because the recipient-level inputs differ per mailbox, placement is a distribution, not a constant. A seed test samples that distribution across providers. This is also why one seed snapshot is noise: our protocol measures over multiple send days before publishing a rate.

Gmail adds a wrinkle: the Promotions and Updates tabs. Tab placement is not spam — the mail is delivered, findable, and opened at meaningful rates — so we count categorized placement as its own state rather than a failure.

Inbox placement and your deliverability

Placement is deliverability, in the end. Authentication records, reputation dashboards and warm-up schedules are all instruments; the inbox rate is the outcome they exist to produce. That is why our rankings sort ESPs by measured placement rather than feature checklists, and why every score we publish carries its test date.

For your own program, placement data changes decisions in ways sender-side metrics cannot. A falling open rate is ambiguous — worse subject lines, list fatigue, or a placement collapse all look the same from the sending side. Seed data disambiguates it. Placement broken out by provider does more: a sender inboxing at Gmail and junking at Outlook has a specific, diagnosable problem, not a general one.

Providers offer partial visibility from their side. Google Postmaster Tools reports domain-level spam rate and reputation for Gmail; Microsoft SNDS shows complaint and trap data for your IPs at Outlook. Neither reports placement directly, but both correlate with it and both are free to register.

Limitations and failure modes

Trusting the delivery number. More placement problems hide behind this than behind anything else: a sender watches "delivered: 99%" for months while placement erodes, and learns of the problem only when revenue reports it. Acceptance metrics cannot detect a placement problem by construction.

Seed lists are a sample, not a census. Seed mailboxes approximate real mailboxes but lack personal engagement history with you. A seed test measures the provider's baseline treatment of your mail; your engaged subscribers typically do somewhat better, your dormant ones worse. Treat seed rates as comparable benchmarks, not as the literal experience of every recipient.

Single-send conclusions. Placement shifts day to day with reputation, volume and content. One test is an anecdote. Trends across repeated sends are data — our protocol requires multiple send days for exactly this reason.

Panel-data "placement" estimates. Some tools infer placement from opens across a user panel. Inferred rates degrade as open tracking degrades (image proxying, privacy features), and panel composition skews consumer. Know which kind of number you are reading.

Chasing the tab. Senders burn effort trying to escape Gmail's Promotions tab into Primary. Marketing mail in Promotions is where Gmail's users expect it; aggressive tricks to escape it read as evasion and risk the spam folder, a strictly worse outcome.

Ignoring the missing. Mail can be accepted and then discarded without appearing anywhere. Silent loss is rare but real, and it is invisible unless your measurement counts it as its own category — ours does.

Related terms

Sender reputation, domain reputation, IP reputation, spam complaint rate, DMARC, email warm-up, list hygiene, SMTP.

Frequently asked questions

What is a good inbox placement rate? There is no universal published benchmark, and any precise industry average you read is a vendor estimate. The practical standard: overwhelmingly inbox plus tabs, spam in low single digits, missing near zero. Our per-service test results show the measured range across real platforms.

Why does my ESP show 99% delivered if my emails go to spam? Because "delivered" means accepted by the receiving server. Spam-folder placement happens after acceptance and is never reported back to the sender. The metric is honest about what it measures; the name misleads.

How can I check my own inbox placement? Send real campaigns to seed mailboxes you control at Gmail, Outlook and Yahoo and record where they land, over several sends. Registering for Google Postmaster Tools and Microsoft SNDS adds provider-side reputation context.

Is Gmail's Promotions tab a placement failure? No. Tab placement is delivered mail, categorized. It underperforms Primary for attention but vastly outperforms spam. We score it as its own state, at partial credit, and recommend senders do the same.

Does inbox placement differ by mailbox provider? Substantially — it is common for a sender to place well at one provider and poorly at another, because each runs its own filtering with its own data. Per-provider breakdown is the most actionable cut of placement data, which is why we publish Gmail, Outlook and Yahoo views separately.

Can any service guarantee inbox placement? No, and the claim is a red flag. Placement is decided by receiving providers, per message, per recipient. Anyone guaranteeing it is selling something the filtering architecture does not permit.

If a placement question brought you here, start where the evidence is: our test methodology explains exactly how the placement numbers on this site are produced, and what they can and cannot tell you.

Sources

• Google Postmaster Tools documentation (domain reputation and spam-rate reporting)
• Google Postmaster guidelines (Email sender guidelines)
• Microsoft SNDS documentation (Smart Network Data Services)
• Yahoo Sender Hub (sender requirements and best practices)