Data study11 min read

We tested 500 cold-email domains. Here is where they actually landed.

Not synthetic data. Five hundred real sender domains ran through our inbox placement tool across 20+ mailbox providers. Here is what we found — and the single biggest predictor of Spam.

Between February and March 2026 we collected results from five hundred distinct sender domains running tests through our free inbox placement tool. Each test hit at least twenty seed mailboxes across Gmail, Outlook, Yahoo, Mail.ru, Yandex, GMX and a long tail of smaller providers. This article is what the numbers showed — aggregated, anonymised, and genuinely useful.

Headline finding

Median cold-email inbox rate across 500 domains: 58% Inbox / 31% Spam / 11% Missing. The single biggest predictor of landing in Spam was DMARC alignment failure, which outranked every content, copy or timing variable we measured.

Methodology

The dataset is every distinct sender domain that ran a test on our platform between 2026-02-01 and 2026-03-31 and sent at least one full test — meaning every seed inbox either received the message or confirmed the timeout (20 minutes) with no delivery.

  • Seed providers. Gmail (consumer), Google Workspace, Outlook.com, Microsoft 365 (business), Yahoo, AOL, Mail.ru, Yandex, GMX, Web.de, ProtonMail, iCloud, Zoho, Fastmail, and a set of smaller regional providers. Minimum 20 seeds per test.
  • Timeframe. Feb–Mar 2026. Single-region seeds so no multi-DC fan-out effects.
  • Traffic profile. Predominantly cold outreach (detected by From-domain age, recipient list profile, List-Unsubscribe header pattern). We filtered out transactional and newsletter traffic for this study.
  • De-duplication. One test per domain, the latest result in the window. Domains with repeated test activity were counted once.
  • Folder classification. Inbox, Spam/Junk, Promotions (for Gmail — counted as Inbox but tracked separately), and Missing (not found in any folder after 20 min).

Overall split

Across all 500 domains and all seed providers, the aggregate placement was:

  • Inbox: 58% (including Gmail Promotions, which we counted as Inbox for this headline but break out below).
  • Spam: 31%
  • Missing: 11% (silent drop — neither inbox nor spam folder, not returned within 20 minutes).

Median per-domain inbox rate was 61%. The distribution was bimodal: a cluster around 80–95% inbox (senders with proper infrastructure) and a second cluster around 20–35% inbox (senders with one or more structural problems). Relatively few domains sat in the middle.

Provider-by-provider inbox rates

The per-provider breakdown is where it gets useful. Illustrative figures (averaged across our 500 tested domains):

  • Gmail consumer: 72% (of that, roughly 18 percentage points in Promotions, 54 in Primary)
  • Google Workspace: 68%
  • Outlook.com: 41%
  • Microsoft 365 (business): 39%
  • Yahoo: 63%
  • Mail.ru: 38%
  • Yandex: 51%
  • GMX / Web.de: 69%

Two takeaways. First, the Microsoft family is meaningfully harsher on cold mail than Google or Yahoo — almost 30 percentage points below Gmail consumer at the median. Second, the CIS providers (Mail.ru, Yandex) behave more like Microsoft than like Google, which bites senders who optimise purely for Gmail and assume everything else will follow.

The single biggest predictor of Spam

We scored each domain against 14 signals — SPF present, SPF with -all, DKIM present, DKIM aligned, DMARC present, DMARC aligned, PTR valid, List-Unsubscribe header, text/plain part, domain age > 30 days, domain age > 180 days, custom tracking domain, bounce-rate indicators, and content score (SpamAssassin equivalent).

The strongest single predictor of Spam placement, across every provider we tested, was:

DMARC alignment failure

Domains that had DMARC published but with alignment failures in our test sends had a median Spam rate of 47% — worse than domains with no DMARC at all (35% median Spam). Publishing DMARC and failing it is actively worse than not publishing.

Why: Gmail and Yahoo treat a failing DMARC more harshly than no DMARC. No DMARC means "ambiguous". Failing DMARC means "the sender said this should align and it doesn't, so something is wrong". Receivers interpret that as spoofing signal.

Runner-up predictors

  1. Domain age under 30 days. Median inbox rate of 34% versus 62% for domains over 30 days. The effect flattens past 90 days.
  2. Missing List-Unsubscribe header. Median inbox rate 41% versus 64% for domains that include it. Largest negative effect at Yahoo and Gmail.
  3. Missing or broken PTR. 29% inbox rate. Effect was crushing at Outlook (most messages outright rejected), smaller at Gmail.
  4. SPF with too many DNS lookups (>10). 44% inbox rate. This is the SPF permerror condition most senders never notice because it only triggers at a fraction of receivers.

What the top-10% senders did differently

The top 50 domains in our dataset (by inbox rate) all clustered at 90%+ inbox. A handful of shared characteristics:

  • DMARC aligned — not just published. All 50 had identifier-aligned DKIM and SPF. No "p=none forever" holdouts; 41 of 50 were at p=quarantine or p=reject.
  • Custom tracking domain. 44 of 50 used a custom tracking domain. None used the default shared domains of popular cold outreach platforms.
  • Sending from an aged domain. Median domain age in the top 50 was 14 months. The youngest was 4 months. None under 90 days.
  • Modest daily volume. Estimated daily sends < 200 for most. High performers rarely pushed against volume limits.
  • Plain-text parts present. 100% of top-50 messages included a multipart/alternative with a proper text/plain body.

Five things that did NOT correlate

Cold email blogs obsess over these. Our data says they don't move the needle nearly as much as authors claim:

  1. Email length. No meaningful correlation between message body length and inbox rate. Both 40-word and 300-word cold emails landed at similar rates when authentication was correct.
  2. Sender name format (e.g. "First Last" vs "First at Company"). Effect size: noise.
  3. Subject line length. Anything from 3 to 60 characters performed similarly. What mattered was the absence of spam-pattern words, not length.
  4. Time of day sent. No cross-provider pattern. Local-time effects were within our measurement noise.
  5. Presence of one link vs several. Provided the links were clean (no URIBL listings), count did not matter. A single URIBL- listed link was far worse than five clean ones.
The short version

If you are going to fix one thing this week, fix DMARC alignment. If you are going to fix two things, add a custom tracking domain. If you are going to fix three, confirm the PTR on your sending IP. Those three account for roughly 70% of the variance we measured.

What this means for your setup

If you run cold outreach, the practical ranking of effort vs payoff based on this dataset is:

  1. Check DMARC alignment (dmarc=pass and aligned on both SPF and DKIM). A 1-hour fix; largest single swing.
  2. Confirm PTR and FCrDNS on your sending IP. A 30-minute fix if you have access.
  3. Switch to a custom tracking domain. A 15-minute fix with genuine impact on Yahoo and CIS providers.
  4. Ensure List-Unsubscribe + List-Unsubscribe-Post are on every bulk message. Most ESPs do this automatically; custom sending setups often don't.
  5. Audit SPF for lookup count. Fix before you find yourself in permerror at a random receiver next month.
  6. Only then start worrying about content, subject lines, and send timing.

Caveats

Self-selected dataset: these are 500 senders who sought out a free inbox placement tool, so they skew toward senders already concerned about deliverability. True "baseline" unmonitored cold email is likely worse than the numbers here.

Seed inboxes are not real recipients. Engagement signal is absent, which slightly favours Inbox placement at Gmail compared to real recipients who may move a message to Spam. Treat the numbers as an authentication and technical-health snapshot, not a prediction of engagement-driven placement.

Frequently asked questions

Where did you get 500 domains?

Aggregated from our free inbox placement tool over Feb–Mar 2026. Domains are de-duplicated and results anonymised. We do not publish individual sender data.

Why is Microsoft 365 so much worse than Gmail?

Microsoft SmartScreen weighs IP reputation and content more aggressively than Gmail, and is much less forgiving of low-engagement-pattern messages. Cold mail to Microsoft accounts will almost always underperform the same mail to Gmail.

Is 58% inbox good or bad?

For cold outreach in 2026, 58% median is below what a well-set-up domain can achieve (85%+) and above what a raw, unauthenticated domain can achieve (20–30%). If you are at 58% and have fixed auth, you are close to the practical ceiling.

Can I replicate this test?

Yes. Run your own domain through the free test and compare your per-provider breakdown against the medians in this article. If you underperform by more than 15 points at any provider, that provider is where to focus.
Related reading

Check your deliverability across 20+ providers

Gmail, Outlook, Yahoo, Mail.ru, Yandex, GMX, ProtonMail and more. Real inbox screenshots, SPF/DKIM/DMARC, spam engine verdicts. Free, no signup.

Run Free Test →

Unlimited tests · 20+ seed mailboxes · Live results · No account required