Cold email deliverability benchmarks 2026: Google Workspace against Microsoft 365
Across 1,160,919 cold emails sent for 45 clients over 90 days, the bounce rate was 2.58% and 14.1% of classified replies were positive. Split by the recipient's email gateway, Google Workspace produced 75.2% of all replies at a 14.31% positive rate, and Microsoft 365 produced 24.2% at 13.70%.
Methodology
The headline contradicts what most outbound teams believe: once a message reaches the inbox, Google and Microsoft recipients behave almost identically. The difference between the two is not what happens after delivery. It is delivery itself.
- Source
- Danish Lead Co's own delivery platform, the system that runs client campaigns day to day. All figures are aggregate and anonymised. No client, campaign or prospect is identifiable.
- Window
- 25 June to 23 September 2026, a 90-day window.
- Population
- Client campaigns only: 45 client companies, 344 campaigns, 1,160,919 emails, 520,706 unique prospects. Danish Lead Co's own prospecting and its sub-brand campaigns are excluded.
- Classification
- Replies are classified by our own inbox system, not by the sending tool. Positive means the prospect asked a real question, asked for a call, or referred us on. Out-of-office is counted separately and never as positive.
- Gateway
- Attribution is taken from the recipient domain's mail exchange records, not from guessing at the domain name. Rows where the recipient address resolved to one of our own sending domains are excluded, which removes 0.20% of replies.
- Reply sample
- 13,722 classified replies after that exclusion.
The gateway table
| Recipient gateway | Replies | Share of replies | Positive | Positive rate |
|---|---|---|---|---|
| Google Workspace | 10,318 | 75.2% | 1,477 | 14.31% |
| Microsoft 365 | 3,314 | 24.2% | 454 | 13.70% |
| Other infrastructure | 90 | 0.7% | 2 | 2.22% |
Other infrastructure is 90 replies, too small to draw anything from, and it is in the table only so the column adds up.
Microsoft 365 is not a quarter of business email. The split is a delivery figure, not a market share figure.
Why teams misread it
Microsoft's filtering is materially more aggressive on cold mail than Google's, so the 75/24 split reflects how many messages reached the inbox in the first place.
Teams read a low Microsoft reply count as "Microsoft recipients are not interested" and rewrite their copy. The copy was never the problem.
What the positive rates show
14.31% against 13.70%, within one percentage point. A prospect on Microsoft 365 who reads the message is very nearly as likely to be interested as one on Google Workspace.
Everything lost is lost before the read. So deliverability work beats copy work on Microsoft-heavy lists: industrial, manufacturing and financial services lists heavily skew that way.
The bounce number, and why it is low
2.58% across 1,160,919 sends. An unverified list routinely bounces above 8%. The difference is verification before sending, every address, every campaign.
It matters beyond the wasted sends: bounce rate is a reputation signal, and the damage from a bad campaign outlives it by months on the same domain.
What we would change first on a Microsoft-heavy list
- Authenticate properly. SPF, DKIM and DMARC aligned on the sending domain, not inherited.
- Lower volume per mailbox and add mailboxes, rather than pushing more through the same ones.
- Warm the domain for longer than feels necessary before the first real campaign.
- Verify every address. 2.58% is achievable and it is mostly this.
- Only then look at the copy.
The rest of the study
The parent report
Every headline figure in one place, with the full methodology and what we deliberately do not report.
Cold email benchmarks
Every denominator published, and how to compare your own tool's numbers to them honestly.
LinkedIn outreach benchmarks
11,701 connection requests across 10 clients. 83.6% of the funnel is lost before a single message is sent.
Who actually replies, by seniority
Founders at 13.51%, VPs at 2.87%, and why the middle of the org chart performs worst.
Frequently asked questions
What is a good cold email bounce rate?
Under 3%. We run 2.58% across 1,160,919 sends with verification on every address. Above 5% will damage the sending domain.
Do Microsoft 365 recipients reply less than Google Workspace?
They produce fewer replies in total, 24.2% against 75.2%, but once they reply they are almost as likely to be positive: 13.70% against 14.31%. The gap is delivery, not interest.
Does Microsoft block cold email entirely?
No, but it filters more aggressively than Google. The practical answer is more sending domains, lower volume per mailbox, and proper authentication.
Should we avoid Microsoft-heavy industries?
No. Manufacturing, logistics and financial services skew Microsoft and are among the most valuable B2B markets. Budget more deliverability work for them.
How often is this updated?
Quarterly, from the same platform. The window is stated on the page so a figure can always be dated.
Want these numbers run on your own list?
Most Microsoft-heavy lists are fixable. If yours is underperforming, the call is where we look at the sending setup rather than the copy.
Source: Danish Lead Co delivery platform / Window: 25 June to 23 September 2026 / Population: 45 client companies, 344 campaigns, 1,160,919 emails, 520,706 unique prospects / Aggregate and anonymised. Danish Lead Co's own campaigns excluded. / Updated quarterly.