Spinning up
Live
+312AI citations this month
Press⌘Kto search
Spinning up
Cold email and LinkedIn outbound benchmarks from inside 30 paid client engagements and 80 control programs. 1.2M sends, 130 sender domains, 90 days.
Section 01 · Headline numbers
Median reply rate on tight ICP
11.4%
vs 3.1% on loose ICP lists in the control set
Inbox placement after proper warmup
78%
vs 47% on cold-launch domains skipping warmup
Lift from AI personalization opener
+62%
Positive reply rate 5.2% to 8.4% on matched ICP
Sent-to-meeting conversion (healthy)
1.5-3%
End to end on tight ICP with reasonable offer
Section 02 · Channel mix
Cold email still dominates volume but LinkedIn matched it on reply rate per message in 2026. The strongest performers ran both in coordinated multi-touch sequences.
Section 03 · Deliverability
Across 130 sender domains in the tracked corpus, the following technical and behavior factors drove inbox placement. Numbers are observed placement rates per cohort.
| Factor | Placement rate |
|---|---|
| Proper 3-6 week warmup | 78% |
| SPF + DKIM + DMARC aligned | 73% |
| Sub-domain (mail.brand.com) sender | 71% |
| Sequence under 4 emails | 66% |
| Reply-to baked into sequence (asks question) | 61% |
| No warmup (cold-launched domain) | 47% |
Section 04 · What gets a reply
Across 380,000 emails with positive replies, the openers clustered into five patterns. Numbers are share of positive replies attributable to each pattern.
Cold opener that mentions a specific system in the recipient's stack and a known failure pattern. Required research, paid off in reply quality.
Opener with a numerical claim narrow enough to be defensible. Generic statistics underperformed by a wide margin.
1-2 sentence cold note with a single yes/no question. The brevity itself was the differentiator.
Real mutual connection mentioned in first line. Synthetic name-drops (we both know X who you don't actually know) got flagged and hurt reply quality.
Reference to a recent funding round, exec hire, or public announcement that triggered the outreach. Highest meeting-conversion of all patterns.
Section 05 · Sequence length
Sequence length is the most over-optimized variable in cold outbound. Longer is not better. Across the corpus, reply rate per email peaked at email 3 and declined sharply after email 5.
| Sequence step | Reply rate | Cumulative meetings |
|---|---|---|
| Email 1 | 5.1% | 47% |
| Email 2 | 4.3% | 75% |
| Email 3 | 3.8% | 91% |
| Email 4 | 1.4% | 97% |
| Email 5 + | 0.7% | 100% |
Section 06 · Methodology
Window. January 1 to March 31, 2026. 90 days of continuous tracking.
Cohort. 30 paid Xpand Media B2B outbound engagements + 80 non-client control programs sampled across SaaS, fintech, AI infrastructure, and professional services.
Volume. 1.2M sends in the window across 130 sending domains. Average daily send volume per domain: 75 messages.
Tools tracked. Instantly, Smartlead, Salesloft, Apollo, Outreach. The corpus skews toward Instantly and Smartlead, which together represent 64% of tracked engagements.
What counts as a reply. Any human response to a sequenced email or LinkedIn DM. Auto-replies and bounces excluded.
Limits. Predominantly English-language outbound to US, EU, MENA, and APAC buyers. Numbers for German, Arabic, and Japanese language outbound carry larger error bars.
Section 07 · Frequently asked
Cold email and cold LinkedIn outreach to companies and contacts the brand has had no prior relationship with. Includes sequence automation via Instantly, Smartlead, Salesloft, and Apollo. Excludes warm referral plays and event-driven follow-up.
Meetings booked is the right end-state metric but lags by 2-6 weeks and depends on calendar friction outside the outbound system. Reply rate is the cleanest signal of campaign quality controllable inside the sequence.
Decisive. Cohorts that skipped warmup had a 47% inbox placement rate against a 78% placement rate for properly warmed cohorts (n=50 sending domains, Q1 2026). The difference shows up in reply rate before it shows up in spam-complaint metrics.
Yes for the opener line. The personalization-AI cohort lifted positive reply rate from 5.2% to 8.4% on tight ICP. Caveat: AI personalization on broken ICP made things worse, not better. Sequence the ICP discipline first.
No. LinkedIn outbound matches cold email reply rate on average and beats it for buyers who actively filter cold email. The dominant pattern in 2026 is multi-touch: LinkedIn connect + cold email + LinkedIn message follow-up.
1.5 to 3% on a tight ICP with reasonable offer. Below 0.8% suggests list quality, deliverability, or offer problems. Above 5% almost always means the list is small and the offer is exceptionally well-matched.
GDPR explicitly permits B2B cold outreach when the messaging is relevant to the recipient's professional role and includes opt-out. CAN-SPAM in the US requires accurate identification and an unsubscribe path. Both let well-run B2B outbound continue legally.
List quality. In the tracked corpus, swapping a 5000-contact loose list for a 800-contact tight ICP list lifted reply rate from 3.1% to 11.4% with the same sequence and sender domain.
30 paid Xpand Media B2B outbound engagements plus 80 control programs, tracked January through March 2026. Total volume in the window was 1.2M sends across 130 sending domains.
Yes. Free to cite under a CC BY 4.0 license. Plain text and BibTeX citation blocks are below.
Section 08 · Cite this report
Plain text
Xpand Media (2026). State of B2B Outbound 2026: Tracked corpus deliverability and reply benchmarks. https://xpandmedia.io/state-of-b2b-outbound-2026
BibTeX
@techreport{xpandmedia_b2b_outbound_2026,
author = {{Xpand Media}},
title = {State of B2B Outbound 2026: Tracked corpus deliverability and reply benchmarks},
institution = {Xpand Media},
year = {2026},
url = {https://xpandmedia.io/state-of-b2b-outbound-2026}
}License: free to cite and quote in commercial and academic contexts. Direct link backs to the report URL are appreciated but not required. The underlying anonymized dataset is available on request.
We run the same tracking corpus on every active outbound client. A 20-minute call gets you a written read on where your reply rate sits against the corpus, what the biggest leak is, and the single highest-leverage move.