Inbox Placement Rates 2026: The Numbers Worth Trusting

As of 2026, the global average inbox placement rate is 87.2%, per Validity's 2026 Email Deliverability Benchmark — and three sources with different methods converge on 13–18% of legitimate commercial email never reaching the inbox, roughly one message in six to one in eight. Inbox placement rate is the share of sent mail that lands in the inbox rather than the spam folder, which makes it stricter and more honest than the "delivery rate" your ESP dashboard shows. Every number on this page carries its source, its date, and a note on how much to trust it — including a section on the famous statistics you should stop citing.

What "inbox placement" actually measures — and the catch behind every number

Delivery rate counts mail the receiving server accepted. Inbox placement counts mail that reached the inbox. The gap between them is the spam folder, and it is large: EmailToolTester's rolling seed tests across 15 ESPs find an average of 10.5% of mail in spam and another 6.4% missing entirely, against an 83.1% inbox average (rolling 2026 data; note the site is affiliate-monetised and the test is seed-based).

Here is the catch: nobody can observe inbox placement directly at scale. Mailbox providers do not publish it. Every published figure comes from one of two proxies:

• Seed lists — test mailboxes the measuring company controls. Their structural flaw, laid out by deliverability practitioner Al Iverson in a July 2026 critique: seed accounts have no engagement history. They never open, click, reply, or rescue mail from spam — and engagement is the dominant modern filtering signal. Seed tests measure precisely the scenario where the strongest positive signal is absent by construction. Add rotating shared IP pools, sample sizes of dozens against providers scoring billions of messages daily, and content drift between test rounds, and Iverson's verdict on cross-ESP placement charts — "inaccurate and misleading" — is hard to argue with.
• Panel data — anonymised data from real mailboxes whose owners installed a monitoring tool. Larger and more realistic, but the panel skews toward whoever installs such tools, and the vendors who run panels (Validity above all) publish no methodology.

So treat every inbox placement statistic, including the ones below, as a directional band measured through a distorting lens — useful for comparison, useless as a precision instrument. If you want ground truth for your own domain at the provider that matters most, Google Postmaster Tools gives you Gmail's own view of your spam rate and authentication — limited, but first-party.

Inbox placement by mailbox provider

The only provider-level breakdown published in 2026 comes from Validity's benchmark. One caveat before the table: the report is form-gated. Validity's own page publishes no figures, no sample size, and no methodology on the open web — only a claim of "trillions of global inbox data points". The numbers below are reproduced from a detailed analyst write-up (Agile Brand Guide, April 2026) whose granularity indicates the author read the full report. That is corroboration, not verification.

Provider	Inbox placement rate	Note
Gmail	89.8%	Best of the majors; Gmail holds 42.9% of the global mailbox market per the same report
Yahoo	87.3%	—
Apple	82.0%	—
Microsoft/Outlook	77.4%	Worst of the majors — nearly 1 in 4 messages misses the inbox

Source: Validity 2026, via Agile Brand Guide, 2026-04-09. Methodology not public.

A version of the Microsoft figure quoting 75.6% circulates in some secondary coverage. The detailed analyst write-up gives 77.4%, and that figure has two independent corroborations — but since the report itself is gated, neither number can be checked at source. "Roughly 77%" is the defensible phrasing.

The Microsoft gap matters more for B2B senders than the headline suggests: the same report puts corporate mail-filter market share at Proofpoint 50.5% and Mimecast 20.7% — 71.2% of corporate mail filtered by two companies, before Microsoft's own filtering even applies. If you sell to businesses, the hardest-to-reach fifth of the market is probably where your buyers live.

Inbox placement by region and country

Same source, same caveat — Validity 2026 via the Agile Brand Guide write-up:

Region / country	Inbox placement rate	Note
Germany	97.5%	Highest measured country
Europe (overall)	91.1%	—
South America	88.0%	Fastest improving, +10.6% year over year
Global average	87.2%	Up 3.7 points year over year
India	71.0%	—
China / Asia	57.9%	Attributed in the report to the Great Firewall

The 40-point spread between Germany and China is the most underappreciated fact in this table: "global average" statistics blend markets whose filtering environments barely resemble each other. If your list is geographically concentrated, the global 87.2% tells you very little.

The cross-check: how much legitimate email misses the inbox

A single vendor's number is a claim. Three numbers from three different methods, customer bases, and commercial interests landing within five points of each other is evidence:

Source	Method	Missed inbox	Date
Validity 2026	Seed + panel network, methodology undisclosed	12.8%	Mar 2026
Sinch Mailgun Email Impact Report	400+ billion sent emails plus a 1,234-respondent survey	18%	Apr 2026
EmailToolTester	Seed-list test, 15 ESPs	16.9%	rolling 2026

The defensible headline: roughly 13–18% of legitimate commercial email never reaches the inbox. Sinch Mailgun deserves a specific credibility note in both directions — its methodology is the best-disclosed in the field (769 of its 1,234 survey respondents are Sinch's own customers, which the company states openly), but that customer base skews transactional, a segment with structurally better deliverability, so its 18% may understate the problem for promotional senders.

For context on why filters are this aggressive: Kaspersky's telemetry put spam at 44.99% of global email traffic in 2025 (published February 2026; denominator is mail seen by Kaspersky's installed base). Mailbox providers are triaging a stream that is nearly half hostile. Some legitimate mail getting caught in that triage is the cost of the system working at all.

Engagement benchmarks that survive scrutiny

Deliverability statistics travel with engagement benchmarks, so here are the ones the sourcing actually supports — and the ones it does not.

Metric	Value	Source	Grade
Gmail spam-complaint threshold	0.3% enforcement / 0.1% target	Google sender guidelines, in force since 2024	Primary — the hardest number on this page
Real-world cross-industry complaint rate	0.02%–0.05%	HubSpot benchmarks, via secondary, 2026	Weak but plausible — roughly 10× below Google's threshold
Apple's share of tracked email opens	62.26%	Litmus Email Analytics, >1bn opens, Jul 2026	Strong primary, large n
Cross-industry average open rate	19.21%	WebFX aggregation of Mailchimp + Campaign Monitor data (30bn+ emails), Dec 2025	Secondary aggregation — see the open-rate warning below
Cross-industry average CTR / CTOR	2.44% / 6.81%	WebFX aggregation, Dec 2025	Same source; clicks are trustworthy where opens are not
Cross-industry average bounce rate	2.48%	WebFX aggregation, Dec 2025	Practitioner consensus treats >2% as concerning, >5% as dangerous
SPF adoption (top-1M domains with MX)	93.8%	RIPE Labs / OpenINTEL, Jul 2026	Independent daily DNS measurement — the best source in this entire field
DMARC records at enforcement	46.9%, falling ~0.44 pts/month	RIPE Labs / OpenINTEL, Jul 2026	Same source; DMARC records are common, DMARC that does anything is a coin flip

The open-rate warning. Since September 2021, Apple Mail Privacy Protection pre-fetches tracking pixels whether or not a human looked, and Apple accounts for 62.26% of tracked opens (Litmus, above). That makes every open-rate benchmark since 2021 structurally inflated by an amount nobody has actually measured — the "+15–20 points" figure you will see repeated everywhere is a practitioner estimate with no primary study behind it. Click rate, click-to-open rate, and reply rate are the engagement metrics still worth benchmarking. If your open rates look great while your clicks are flat, you are probably measuring Apple's servers. What to do about weak real engagement is covered in how to improve email deliverability.

The unsubscribe rate nearly tripled — and almost nobody is writing about it

Buried in Zeta Global's quarterly benchmark reports is, to our reading, the most interesting deliverability datapoint of the last two years: the platform's median unsubscribe rate rose from 0.08% in 2024 to 0.22% in Q1 2026 — a 2.75× increase (baseline in the Q4 2024 report PDF; Q1 2026 figure covered by independent analyst Chad S. White). That movement dwarfs every shift in open and click data over the same period.

The timing is the story. Gmail and Yahoo's bulk-sender requirements mandated one-click unsubscribe (RFC 8058) from February 2024 — the full timeline is on our email deliverability changes page. The plausible reading: the rules worked. Unsubscribing got easier, so people do it. That is correlation, honestly labelled — no controlled analysis has isolated the cause — but no competing explanation fits the timing as well.

Two second-order lessons hide in the same number:

1. Rising unsubscribes are arguably good for deliverability. Every one-click unsubscribe is someone leaving quietly instead of hitting "report spam" — and the spam-complaint rate, not the unsubscribe rate, is what Google enforces at 0.3%.
2. Mean versus median matters. Zeta's median is 0.22% while the WebFX cross-industry mean is 0.89% — a fourfold gap that tells you the unsubscribe distribution has a brutal tail of senders churning their lists. Any benchmark quoting only a mean is hiding that tail.

Statistics to distrust

This section exists because the most-quoted numbers in email marketing are the least rigorous. Before you put any of these in a deck, know where they come from:

The claim	The reality
"Email returns $42 for every $1"	UK DMA Marketer Email Tracker, 2019 — the figure was £42 per £1, UK-only, self-reported. Routinely cited as dollars, current, and global: all three wrong.
"Email returns $36 for every $1"	Litmus, 2020 — a self-reported marketer survey, not measured revenue. Respondents typically count only ESP fees as cost; analyses including labour and data put realistic ROI nearer $15–25. Litmus itself now publishes a distribution rather than the single number — a tacit admission.
"22.5% of your email list decays every year"	MarketingSherpa research, pre-2013, measuring B2B contact-data decay (job changes), annualised by HubSpot and applied ever since to consumer lists it never measured. Modern verifier datasets suggest 22–30%/year — but they measure lists bad enough that someone paid to clean them.
"Undelivered email costs US businesses $59.5bn/year"	A derived estimate multiplying three unrelated benchmarks (Radicati volume × Klaviyo ecommerce revenue-per-email × Validity's missed-inbox share), assuming spam-foldered mail would have converted at average rates. Credit to Mailtrap for showing the arithmetic; it is still not a measurement.
"Apple MPP inflates opens by 15–20 points"	Universally repeated, backed by no controlled study. Cite Litmus's measured 62.26% Apple open share instead.
Any ESP deliverability ranking	Seed accounts have no engagement history — the dominant filtering signal — and the ranking pages are affiliate-monetised for the products they rank. Iverson's summary: "It's you, not the platform."

And the biggest gap of all: no independent warmup study exists

We looked specifically, because we sell warmup software and wanted the independent number. There is no independent, peer-reviewed, or third-party-audited study of email domain warmup anywhere. Every warmup statistic in circulation is published by a company selling warmup software — "90%+ deliverability after 2 weeks" (Warmy), "un-warmed domains start at 40–60% inbox placement" (inboxwarm.ai) — and none publishes a control group, a sample size, or a per-provider breakdown. Worse, warmup networks largely measure mail sent into their own network of accounts, which measures the network, not the inbox. WarmEnvelopes takes a different approach — warmup mail goes to inboxes you control, with every send logged — but that changes the mechanism, not the evidence gap: no vendor's numbers, ours included, substitute for the study nobody has run.

What is credible on new domains: Microsoft's own deliverability guidance — the only mailbox-provider figure on the subject — that a new domain takes 4–8 weeks to reach maximum deliverability, and Spamhaus's published best practice on new-domain sending, which comes from a blocklist operator with no warmup product to sell. Those two, plus the documented fact that enterprise filters weight domain age, are the honest foundation of the warmup case.

How to read any deliverability statistic

The pattern behind everything above compresses to six questions. As of August 2026, most published email statistics fail at least two:

1. Who measured it, and do they sell the solution? Vendor telemetry can still be useful (Zeta, Litmus, Sinch) — vendor marketing claims with no method rarely are.
2. What is the denominator? "Spam rate" of inbox-delivered mail, "opens" including Apple's pre-fetches, and "deliverability" of seed lists are three different quantities wearing one name each.
3. Is the methodology public? The two most-cited reports in the field (Validity, Valimail) are form-gated with no sample size disclosed. The best source (RIPE Labs / OpenINTEL) is free and open. Openness and quality correlate almost perfectly here.
4. Mean or median? Skewed distributions make means misleading — the 0.89% vs 0.22% unsubscribe gap is the live example.
5. Does the time series cross September 2021? If it charts open rates across the Apple MPP launch, the chart is invalid, however pretty.
6. Has the number been re-dated? The £42 ROI figure is from 2019; the 22.5% decay figure predates 2013. Age laundering is the most common failure on this list.