Warm Intro vs Cold Email: 1 in 10 Cheques Is Cold

Of every ten cheques venture capitalists write, about one starts from a cold email. The other nine start from networks, referrals, and the investor's own scouting.

That share is not an opinion. It comes from a peer-reviewed survey of 885 venture capitalists at 678 firms (Gompers, Gornall, Kaplan and Strebulaev, published in the Journal of Financial Economics, 2020). Roughly 30 percent of investments trace to professional networks, about 30 percent to the investor's own proactive search, 20 percent to referrals from other investors, 8 percent to referrals from portfolio companies, and about 10 percent to cold inbound from founders.

Most fundraising tooling is built for the 10 percent channel. This post is about the 90, and about what the 10 is still good for.

What we set out to change

A founder raising for the first time runs the process they can see, and the process they can see is a numbers game: build a list, write a template, send volume. The base rate underneath that plan is severe. The same survey shows a venture firm screens about 200 companies a year and makes about 4 investments, a 2 percent yield before any relationship exists. Practitioner data adds the texture: top partners receive 50 to 200 unsolicited pitches a week and read the first two sentences.

The usual answer is to optimize the email. We have written about what that email costs, and about why most tools only hand you the list in the investor outreach execution challenge (both on the Raisy blog). Copy optimization is real, and its measured sizes are further down. But it is the smaller term, and treating it as the main term is how a founder spends six weeks tuning a channel that starts one cheque in ten.

The idea / how Raisy fixes it

A raise is a networking exercise before it is a copywriting exercise. So Raisy builds warm paths first and sends cold email second.

The core finding, across every dataset we trust: channel beats message by about an order of magnitude. DocSend's 2025 synthesis puts a warm introduction at roughly 13 times the fundraise success rate of cold outreach. The best-measured copy improvements (real personalization, one follow-up, tighter length) land between 1.3 and 2.4 times. Ten times against two times. You tune the big term first.

Warmth also changes what happens to the deck itself.

How it works (the technical layer)

• Grade the introducer, not just the intro. A weak introduction behaves like cold email. In the VCWiz field experiment (arXiv:1806.03241), 301 facilitated, low-context intro requests produced 19 responses (6.3 percent), 5 affirmative replies, 1 connection, and zero investments. The multiplier lives in the quality of the person making the introduction, so we model introducer strength as a graded gate, roughly 1 to 13, rather than as a binary.
• Map the second degree. The path from a founder to a partner usually exists within two hops: a portfolio founder, a shared advisor, a co-investor from a previous round. Raisy surfaces these paths per target from its entity graph instead of leaving the founder to notice them.
• Earn the intro back with research. Reading a fund's portfolio and speaking to its companies first turns a cold request into a warm context. We call the pattern investor reverse diligence, and it is mechanical: pick the target, map the portfolio, talk to the founders inside it, then ask for the introduction with evidence that you did the work.
• Send the deck where the team slide leads. In an AngelList randomized experiment (Bernstein, Korteweg and Laws, Journal of Finance, 2017), early-stage investors responded to founding-team information, while traction numbers and lead-investor names showed no measurable effect. At this stage the team slide is the document; the warm intro is what gets it read.
• Respect the reputation premium. Founders accept 10 to 14 percent lower pre-money valuations from high-reputation firms, and offers from reputable firms are about three times more likely to be accepted (Hsu, Journal of Finance, 2004). The channel premium is priced on both sides of the table, which is why the warm path stays worth the work even when it is slower.
• Run cold as the precise floor, not the strategy. Cold is what fills the top of a warm graph, and precision is what makes it survivable: fit-targeted, verified, signal-specific lists return 5 to 15 percent replies, where generic blasts return under 1.5 percent.

What the user actually sees

In the product this order is visible, not just advised. For each investor target, Raisy shows the path first: who can introduce you, through which node, with the research that makes the ask reasonable attached. Cold outreach runs underneath as the volume floor, with every email researched, approved by you, and logged. Nothing leaves without your approval, and the funnel reports both channels separately, because a reply earned warm and a reply earned cold carry different weight downstream.

The cold levers, measured

For the days when cold is the only door available, these are the sizes of the levers that still work, and one that stopped working.

One follow-up lifts replies by 65.8 percent, and 42 percent of all replies arrive after the first send (Backlinko/Pitchbox, 12 million emails). Personalizing the subject and body with something specific lifts replies by about 142 percent combined, against 7 to 9 percent for basic mail-merge. The response peak sits at 75 to 100 words (51 percent response) at a third-grade reading level (36 percent above college-level writing) (Boomerang, 40 million emails). And one popular trick is finished: first-name merge fields lifted opens 20 percent in a 2016 randomized trial and nothing in the 2024 replication. Tactics have half-lives. Channels compound.

Download every figure in this post (CSV, 30 rows, each with its source and confidence tier)

Why this matters beyond convenience

There is a comforting myth in fundraising advice: that the playing field is level if your email is good enough. The data says the field is made of relationships, and that relationships can be manufactured deliberately, with research and patience, by founders who have no network today. That is the actual levelling. A founder who treats introductions as engineering rather than luck can walk the same 58-contact, 40-meeting, twelve-and-a-half-week path DocSend measures for successful seed closes, and arrive warm at the meetings that decide it.

What this post does not answer

• Whether the 13x figure is exact. It is platform data, published as marketing research rather than peer review. The direction is corroborated by the peer-reviewed deal-flow shares; the exact multiplier is medium confidence, and we label it that way wherever we use it.
• Whether warm paths exist for every founder. Some sectors and geographies have thinner graphs. Syndication platforms cut the geographic hub premium by about a third (NBER WP 24777), and a third is not all of it.
• Whether a warm intro guarantees anything. It buys a reading, not a cheque. The meeting still has to be earned, and the company still has to clear the investor's own bar.
• What your personal funnel looks like. These are population numbers. Your reply rates will differ, which is why the product reports your own funnel back to you rather than asking you to plan against these priors.

A note on confidence

The deal-flow shares, the screening base rate, the AngelList team-signal experiment and the Hsu reputation premium are peer-reviewed, so we treat them as high confidence. The DocSend figures (the 13x multiplier, the review times, the conversion rates, the 58/40/12.5 close budget) are primary platform data published as marketing research, so medium. The cold-email lever sizes come from large vendor datasets of 12 to 40 million emails, where the direction is robust across independent datasets and the magnitudes vary with methodology, so also medium. The VCWiz experiment is a published field study with a small funnel, so medium-high. Where two sources disagree we say so in the figure note rather than averaging them silently.

If you are raising, start from your warm paths rather than your email template. See yours mapped at raisy.team