There is no single cold email inbox placement rate. It changes with the mailbox provider, with how much you send, and with who ran the test.
Recent numbers split by provider come from GlockApps, a company that sells deliverability testing. Across "thousands of GlockApps tests" in the first half of 2026, 49% of test emails reached the inbox.
By provider: Gmail 62.4%, Yahoo 48.9%, AOL 48.7%, Hotmail 36.4% and Outlook 34.7% (GlockApps H1 2026 vs H1 2025 benchmark). That is vendor data from seed tests run on its own platform, with no split between cold and marketing email, so treat the numbers as directional.
For senders of up to 10,000 emails a month, Outlook and Hotmail placement stayed below 18%.
The sender guidelines of Google, Yahoo and Microsoft give no placement rates. They set rules: authentication records, a spam complaint ceiling of 0.3% at Google and Yahoo, and a 5,000-a-day threshold at Google and Microsoft.
Rules and limits below were checked in September 2026.
What inbox placement measures, and what it does not
Several numbers get mixed up in deliverability talk, and only one of them tells you whether a prospect could see your email.
| Metric | What it counts | Who can see it |
|---|---|---|
| Bounce rate | Emails the receiving server refused, as a share of emails sent | You, in your sending tool |
| Delivery rate | Emails the receiving server accepted, which is 100% minus the bounce rate | You, in your sending tool |
| Unsubscribe rate | Recipients who opted out, as a share of emails sent | You, in your sending tool |
| Spam complaint rate | Messages recipients reported as spam, as the provider calculates it | The provider. Google shows it in Postmaster Tools, Yahoo sends reports through its Complaint Feedback Loop |
| Inbox placement rate | Test emails that landed in the inbox, as a share of test emails sent to seed mailboxes | Not part of standard sending stats. A seed test estimates it |
A seed test means sending your real email to a set of test mailboxes at each provider and checking where it landed: inbox, spam, or nowhere.
A bounce announces itself. A spam-folder landing does not: the server accepted the email, your tool logs it as sent, and it sits in the spam folder instead of the inbox.
That is why one agency's 2026 analysis of its own 2025 client sends, which reports bounce and unsubscribe rates by month, country and industry, answers a different question. It shows whether mail was accepted, not where it landed.
Inbox placement by provider and monthly volume
GlockApps splits its 2026 results two ways: by mailbox provider and by the sender's monthly volume. Every cell is the share of test emails that reached the inbox.
| Provider | All volumes | Up to 10K a month | 10K to 50K a month | 50K to 200K a month |
|---|---|---|---|---|
| Gmail (6+ months) | 62.40% | 50.58% | 67.74% | 71.15% |
| Yahoo | 48.87% | 56.93% | 34.92% | 52.13% |
| AOL | 48.71% | 55.24% | 35.60% | 51.75% |
| Hotmail | 36.35% | 17.88% | 34.70% | 41.07% |
| Outlook | 34.67% | 16.88% | 33.53% | 39.06% |
| All providers | 49.00% | 41.27% | 45.20% | 54.31% |
Source: GlockApps H1 2026 vs H1 2025 benchmark, checked September 2026. Vendor data from seed tests run on its own platform, with no stated sample size or seed setup. GlockApps also reports a "Gmail (4+ years)" row, within about 3 points of this one in every tier.
Four things stand out.
Gmail is the easiest of the five once volume passes 10,000 a month. It climbs from 50.6% in the lowest tier to 67.7% and then about 71%. In the lowest tier, Yahoo and AOL edge ahead of it.
Microsoft's two consumer brands sit at the bottom of every tier. For senders of up to 10,000 a month, fewer than 1 in 5 test emails reached an Outlook or Hotmail inbox.
Yahoo and AOL move together. Yahoo's sender rules cover every domain hosted by Yahoo Mail, AOL included, and the two rows track each other within about 2 points across the table. Both dip to about 35% in the middle tier.
The picture moves within a year. The all-provider average rose from 42.98% in H1 2025 to 49.00% in H1 2026. In the lowest tier, Yahoo jumped from 39.48% to 56.93% while Gmail slipped from 57.78% to 50.58%. A benchmark from last year can describe a different market.
One limit. These are consumer mailbox brands. B2B prospects read cold email at their work address, and that company domain may run on Google, on Microsoft or on a regional host missing from the table. The rows show which consumer brand filtered harder in these tests. They cannot predict your own list.
What the provider mix does to your average
Take one real client month from my walkthrough of a cold email sending setup: 15,000 emails over the whole sequence from 20 mailboxes, or about 25 a day per mailbox. At 15,000 a month, a program like that sits in GlockApps' 10K to 50K tier: Gmail 67.74%, Outlook 33.53%.
Now run plain arithmetic on those two averages. The table is an illustration on test data, separate from that campaign's results.
| Who the 15,000 emails go to | Emails reaching the inbox | Share of all 15,000 |
|---|---|---|
| 80% Gmail-type inboxes, 20% Microsoft | about 9,130 | about 61% |
| Even split | about 7,600 | about 51% |
| 20% Gmail-type inboxes, 80% Microsoft | about 6,060 | about 40% |
In this arithmetic, the copy, setup and volume stay fixed. The provider mix of the list alone moves the inbox share by about 20 points.
It also changes what your reply rate means. A reply rate calculated on all 15,000 emails is really calculated on the part that reached the inbox. If 40% reached the inbox, the copy was judged on fewer than half the emails.
You can find your own mix before you send. A company's MX records show whether its mail runs through Google, Microsoft or someone else, and any DNS lookup tool will show them.
What Google, Yahoo and Microsoft officially require
The providers enforce their sending rules through spam filtering, deferral or rejection. The sources are Google's email sender guidelines, Yahoo's sender best practices and sender FAQ, and Microsoft's Outlook.com announcement for high-volume senders.
| Rule | Gmail | Yahoo (and AOL) | Outlook.com (Hotmail, Live) |
|---|---|---|---|
| Bulk threshold | Above 5,000 messages daily to personal Gmail accounts | "We will not specify a volume threshold" | Over 5,000 emails a day from one domain |
| Every sender | SPF or DKIM; TLS; PTR (reverse DNS) plus forward DNS records; spam under 0.3% | SPF or DKIM; PTR (reverse DNS) plus forward DNS records; spam under 0.3% | Enforcement targets the 5,000-a-day senders first; Microsoft says every sender benefits from the same records |
| Bulk senders add | SPF and DKIM, DMARC (policy can be none) aligned with the From domain, one-click unsubscribe on marketing mail | SPF and DKIM, DMARC of at least p=none that passes, From aligned, one-click unsubscribe, opt-outs honored within 2 days | SPF and DKIM must pass, DMARC of at least p=none aligned with SPF or DKIM |
| Spam rate line | Aim below 0.10%, never reach 0.30% | Below 0.3% | None stated; list hygiene and working unsubscribe links recommended |
| If you fail | Delivery can fail, or mail can land in spam | Mail from domains with high complaint rates may be deferred | Rejected with error 550 5.7.515 since May 5, 2025 |
| Sender tool | Postmaster Tools | Sender Hub and its Complaint Feedback Loop | Outlook.com Postmaster site |
Checked September 2026. These are platform rules. This is not legal advice.
Spread cold sending over many domains and each one stays far below 5,000 a day, so the bulk rows may never formally apply to you. The every-sender row still does. Google's all-sender rules include the 0.3% spam line, and Yahoo does not name a volume threshold for its bulk rules at all.
The spam line is tighter than it looks. Google's 0.10% target works out to one report for every 1,000 messages counted. On cold email, where nobody asked to hear from you, that line is the one to watch.
Why the published numbers disagree
Search for an inbox placement benchmark and you will find figures from under 50% to over 80%. They are measuring different things.
| Source | Headline number | What was tested | Counts as inbox |
|---|---|---|---|
| GlockApps, H1 2026 | 49% across providers | Seed tests on its platform, split by provider and monthly volume | Inbox |
| EmailToolTester, January 2024 round | 83.1% average | Email marketing platforms on shared IPs, about 80 seed addresses at 28 ISPs, SPF, DKIM and DMARC set up | Inbox, including Gmail's Promotions tab |
| One agency's study of its own 2025 client campaigns | Bounce and unsubscribe rates | Mail refused or unsubscribed | Placement not measured |
GlockApps sells the testing tool behind its numbers. EmailToolTester is a software review site, and its seed tests measure the email marketing platforms it reviews.
Different senders, different years, different definitions of the inbox. These numbers cannot be averaged into one market figure. When someone quotes you a placement rate, ask three questions: who sent the mail, when, and whether Promotions counted.
What we control before any placement test
This page compiles other people's tests and the providers' own rules. Smirnov Consulting Group is a Prague-based B2B outbound lead generation agency that runs cold email and LinkedIn campaigns for founder-led B2B companies and books qualified sales calls.
On those campaigns, the things we control come before any placement number.
Authentication goes first. In a 2023 post walking through our process, the reminder was plain: "don't forget to set up SPF, DKIM and DMARC records to avoid spam traps." On every sending domain, before any volume.
Then bounces. Google and Yahoo draw their public line on complaints, at 0.3%. On our own sends we watch a different number first, bounce rate: kept under 0.5%, and only verified addresses get sent to.
The reason is in the checks I run on a lead list: "a flagged domain means your next perfect campaign lands in spam for everybody on the list." Once a domain is already listed, each blocklist has its own removal route.
Then spread. A serious setup runs on 10 to 20 domains with 20 to 40 mailboxes, more for a big company, and each new mailbox warms up for 10 to 14 days before it touches a prospect.
Before you trust a placement number
- Check the date and the tier. A figure from last year, or from a sender ten times your size, describes someone else's situation.
- Ask what counted as inbox. If Promotions counted, the number can look better than a Primary-tab result.
- Split your list by provider. Look up MX records and judge Microsoft-hosted prospects on their own numbers.
- Authenticate every sending domain. SPF, DKIM and DMARC of at least p=none, aligned with the From domain. That covers the authentication part of the bulk rules at all three providers, even if you never reach the volume.
- Watch complaints where the providers show them. Postmaster Tools for Gmail, the Complaint Feedback Loop for Yahoo and AOL.
- Keep bounces under 0.5%. Verify every address before it enters a campaign.
- Run your own seed test before you scale. Read the result by provider, and send more only once the weakest provider is where you can live with it.
Questions founders ask about inbox placement
What is a good cold email inbox placement rate?
In GlockApps' H1 2026 vendor tests, the average was 49%: Gmail 62.4%, Outlook 34.7%. Compare your own seed-test result with the row for your volume tier rather than the headline average. Judge Gmail and Microsoft inboxes separately, because one blended number hides the gap.
Does Google or Microsoft publish inbox placement rates?
Not in their sender guidelines. Google, Yahoo and Microsoft publish requirements and tools: authentication rules, spam rate lines and a sender dashboard. Provider-level placement rates like the ones on this page come from third-party seed tests.
How is inbox placement different from bounce rate?
A bounce is mail the receiving server refused, and your sending tool shows it. Placement is where accepted mail landed, inbox or spam, and a seed test estimates it. A low bounce rate says nothing about the spam folder.
Do the sender rules apply below 5,000 emails a day?
Partly. Google's all-sender rules apply at any volume: SPF or DKIM authentication, and spam kept under 0.3%. Yahoo does not publish a volume threshold for its bulk rules. Microsoft enforces at 5,000 a day and says smaller senders benefit from the same records.