Cold vs warm LinkedIn outreach: what 20 million messages tell us

20 million outreach attempts. Here's what the data says.

The cold versus warm outreach debate usually goes like this: someone says cold outreach is dead, someone else says it still works, and both cite their own experience as evidence.

Two large analyses settle it. Expandi's platform analyzed over 20 million outreach attempts across roughly 13,000 unique LinkedIn accounts throughout 2024. Belkins' 2026 State of LinkedIn Outreach study went further, pairing its own 14,077 contact records with 15.1 million Expandi touchpoints from 2025. Worth being precise about the overlap: Belkins' touchpoints come from the same platform, so these are two reads on one pipe across different years, not 35 million independent data points. Either way, both are large enough to draw real conclusions.

The headline number from Belkins' 2026 study: warm campaigns (messaging existing connections) achieve a 12.2% total reply rate. Cold connection campaigns get 7.9%. That's a 54% advantage for warm outreach.

But the most interesting findings aren't in the cold-versus-warm split. They're in the combinations.

The numbers that matter

Reply rates by campaign type

These four figures all come from the Belkins 2026 study, so they share one sample and one denominator and can be read against each other directly.

The gap between the best approach (messenger, 12.2%) and the worst (builder, 5.4%) is more than 2x. Campaign type matters more than most people's messaging optimization.

Expandi's own H1 2026 report ranks campaign types the same way on a different and more recent sample, at consistently higher absolute numbers: messenger campaigns to existing connections at 16.86%, event campaigns at 14.21%, campaigns aimed at people who viewed your profile at 13.4%, and template-based campaigns at 8.62%, against a 10.3% LinkedIn-wide average. From the earlier 2024 Expandi dataset, paid InMail campaigns averaged 6.38%. Do not compare any of those figures line by line against the Belkins ones above; the samples and periods differ. Read them as independent studies agreeing on the ordering, which is the part that should drive your decision.

The action combination that doubles results

The single most important finding in the Belkins data: combining a direct message with a profile visit before sending produces an 11.87% reply rate, the best two-action combination in the study. Sending a direct message alone gets 4.88%.

That's a 143% improvement from one additional action.

Profile visits create a warm signal. The prospect sees your name and headline in their "Who's Viewed Your Profile" section. When your message arrives, you're not a stranger - you're someone who already showed interest. This is how networking actually works: you notice someone, then approach them.

The cumulative effect of nurturing

Belkins tracked how reply rates change as you add more nurturing actions before outreach:

• 1 nurturing action: 1.07% reply
• 2 nurturing actions: 2.53% reply
• 3 nurturing actions: 3.78% reply
• 5 nurturing actions: 5.26% reply

Each additional touchpoint (profile visit, post like, comment engagement) adds roughly 1 percentage point to reply rates. Five nurturing actions before outreach generates 5x the replies of a single cold touch.

What personalization actually moves

The data on personalization is surprisingly nuanced.

Connection acceptance: notes barely matter

Prospeo's data, cited in Cleverly's April 2026 benchmark set, found that connection requests with a personalized message get 26.42% acceptance. Without a message: 26.37%. Statistically identical. Belkins' 2026 study found the same shape on its own sample, with the blank request actually edging ahead: 27.6% without a note against 25.3% with one.

This challenges the conventional wisdom that you should always include a note. On LinkedIn, acceptance seems to be driven more by your profile (headline, photo, mutual connections) than by the note, a pattern our LinkedIn connection acceptance rate benchmarks break down in more detail. Free accounts are now limited to roughly 5 noted requests per month anyway, making this largely a non-issue for high-volume senders.

Reply rates: personalization matters a lot

While notes don't affect acceptance, they significantly impact what happens after (Prospeo, via Cleverly):

• Personalized first message: 9.36% reply rate
• Non-personalized: 5.44% reply rate

That's a 72% improvement, and Belkins measured the same direction on its own sample: 8.2% reply for noted requests against 5.3% for bare ones. The takeaway: save your personalization effort for after the connection is made, not for the connection request itself.

AI personalization: the surprising results

In the Belkins data, AI-assisted messages outperform non-AI messages on the first touch (4.19% vs 2.60%, a 61% lift). But on follow-ups, non-AI messages perform slightly better (3.91% vs 3.48%).

The likely explanation: AI is good at initial personalization (reading profiles, referencing posts) but produces follow-ups that feel less natural than human-written ones. Total campaign performance with AI: 7.66%. Without: 6.50%. AI adds about 18% lift overall - helpful but not transformative.

When cold still wins

The data doesn't say warm always beats cold. It says warm gets higher reply rates on average. But there are scenarios where cold outreach is the right approach.

When your TAM is small

If you sell to a niche with 200 target companies and you've already connected with most decision-makers, there's no warm pool to fish from. Cold outreach to the remaining prospects is your only option.

When speed matters more than conversion rate

A company launching a new product into an unfamiliar market needs awareness fast. Cold outreach at 7.9% reply rate with 100 daily touches beats warm outreach at 12.2% with 20 daily touches - at least in the short term.

When you have no content engine

Warm outreach depends on creating warmth - publishing content, engaging with posts, building visibility. If you're starting from zero content presence on LinkedIn, cold outreach gets you into conversations while you build your warm engine.

When warm wins decisively

Enterprise and high-ACV sales

When each deal is worth $50K+, the difference between 7.9% and 12.2% reply rates compounds massively through your pipeline. If you need 10 meetings to close 1 deal, warm outreach gets you there with roughly 82 outreach attempts. Cold takes 127. Over a year, that's hundreds of hours saved.

Saturated markets

If your prospects receive 10+ sales messages per week (common for VP Sales, CMOs, and Heads of Growth), cold outreach competes with everyone else in their inbox. Warm outreach - where they've already seen your name through profile visits, post engagement, or mutual content - cuts through because you're not a stranger.

The data confirms this: the most-prospected desks, marketing and finance, sit at the bottom below 6%, while HR and talent acquisition top the list at 10.9% (Belkins 2026). Sales people are numb to sales outreach. You need warmth to break through.

Long-term relationship building

Warm outreach builds compound advantage. Every post like, every comment, every profile visit creates a small signal. Over weeks, a prospect who's seen your name five times is dramatically more likely to respond than one who's never heard of you.

The hybrid approach (what actually works)

The best-performing teams in the data don't choose cold or warm. They build systems that create warmth before cold outreach happens.

The sequence that works

1. Week 1-2: Identify prospects through buying intent signals (engagement, job changes, hiring)
2. Day 1: Visit their profile (creates "who viewed" notification)
3. Day 2-3: Like or comment on one of their posts (second touch)
4. Day 4-5: Send connection request (without note)
5. On accept: Send personalized message referencing their recent activity
6. Day 7-10: Follow up if no reply

This isn't warm outreach or cold outreach. It's a warm-up sequence that turns cold prospects into warm ones before you ever send a message. The Belkins data shows this combination (multiple nurturing actions + personalized message) can achieve 10%+ reply rates consistently.

Tools like BeReach help you run this entire sequence. The BeReach agent helps you find the right prospects, warm them up over a few days, and reach out with a personalized message when they're ready - all at a respectful pace. The result is more replies and more booked meetings from the same list. BeReach plans start at EUR 99 a month flat for two connected accounts, with no per-seat multiplication on top.

Timing: what the data says about when to reach out

Day of week: the spread across weekdays is small enough that it is not worth engineering your calendar around. Weekdays edge out weekends, and that is about all the day-of-week data supports. Don't overthink it.

Month: the intuition that January is a fresh-start goldmine is wrong. Expandi's 2026 report puts January at a 6.47% reply rate, its weakest month, and January connection approvals at 21.45% against an annual average of 29.61%, a 33% drop. Everyone else is also emailing in January. Plan around the dip rather than betting on it.

Geographic performance: in Belkins' 2026 country data, the Netherlands leads at 10.5%, with Denmark and Belgium close behind around 9-10%, then Canada at 7.1%, Australia at 6.4%, the UK at 6.1% and the United States at 6.0%. The regional spread is real but smaller than the gap between campaign types.

Frequently asked questions