Tracing the Warm Path Into a Cold Target Account

You have one cold logo you need to get into, and you want to know who in your company's network - or your own - can hand you a credible introduction to the right buyer. This guide is for founders selling their own product, account executives, SDR leads, and partnerships teams. It carries one cold account through the whole job, with the real intermediate counts and the wrong turns, so you finish with either a ranked list of warm-intro paths or a clean decision to go cold.

Most of what ranks for this job stops at "map your network." That is the easy 20%. The hard part is scoring the paths honestly, verifying the top tie before you spend a connector's capital, and knowing the threshold below which you should stop looking and route the account to cold outreach. That is what I work through here.

Why a warm path is worth the effort at all

A warm introduction is worth the extra hours because the reply and conversion gap over cold is enormous, not marginal. Warm introductions produce reply rates of 20-40% against 1-3% for cold email, and meeting show rates of 85-95% versus 50-60% for cold-sourced meetings. That is the whole argument for spending an afternoon tracing a path rather than firing a sequence.

The published benchmarks line up in one direction:

Referral leads convert at 11%, the highest lead-to-customer rate of any channel, against 0.2-2% for cold sources. B2B companies with referral programs report 71% higher conversion and 69% faster close time per Heinz Marketing research. And 84% of B2B decision-makers begin the buying process with a referral or peer recommendation, so a warm path is not just a better opener - it matches how the buyer already wants to buy.

There is a supply problem hiding under this. Willingness is not the bottleneck. In one study, 83% of satisfied customers said they were willing to refer, but only 29% were ever asked. The paths mostly exist and the people mostly say yes. The work is finding them and asking well.

The worked example: one cold logo, start to finish

Here is the case I will carry the whole way through. The account is a mid-market software company I will call the target. I have no relationship there. My job is to produce a ranked list of warm-intro paths, each with a named connector, the second-degree buyer they reach, the evidence of relationship strength, and a drafted intro-request - or to confirm no warm path exists and route to cold.

The first fork is who to aim at, and this is where most reps aim too high. The instinct is to target the CRO. The data says target the Director.

In Refolk's index of professional profiles, US Director-level revenue buyers outnumber CRO and VP of Sales roughly 4.8x (77,377 versus 16,200). Fewer senior seats means fewer nodes for any tie to land on, so a warm path to a Director is materially more likely to exist - and that person is very often the real evaluator anyway. So the named buyer I write down is the Director of Sales, with the VP of Sales as backup.

Geography compresses the pool further. The same senior-buyer titles are about 17.8x larger in the US than the UK.

For a UK target, an individual rep will far more often hit the no-warm-path case, which is exactly when pooling the whole company's network stops being optional. My example target is US-based, so the buyer node is relatively findable - but I will still pool the team.

Enumerating the connection categories

Before scoring anything, list every first-degree person who touches the account, category by category. The categories that practitioners and vendors converge on are the same short list every time.

• Current colleagues who know someone at the target.
• Former colleagues who now work at the target, or know someone there.
• Current and former customers connected into the account.
• Investors on your cap table with lines into the target's leadership.
• Advisors and board members with the same.
• Partners in your ecosystem who sell alongside or into the account.

One sub-category deserves its own line because it is the warmest path into a net-new logo: the recent job-changer. When someone who already knows you moves into a target account, especially in their first 90 days, that is often the best entry point you will get. Check who joined the target recently and previously overlapped with someone on your team.

For my example, the raw list came back with six candidate paths: two former colleagues now inside the account, one current customer who used to work there, one investor connected to the VP of Sales, one LinkedIn 1st-degree link a teammate had, and one recent job-changer who joined the target's revenue org 40 days ago.

Notice the collapse. Six raw paths became one I would actually use. That collapse is not failure; it is the work doing its job.

Scoring each path by relationship strength

Rank each path by three signal types plus tenure overlap, and put the strongest reciprocal, most recent tie on top. The three signals that drive the score are recency (when was the last meaningful interaction), frequency (how often do they interact), and network depth (how many shared connections exist). Add tenure overlap, which tells a real relationship apart from a forgotten connection.

There is a published benchmark for the top of the range. A contact with 12 email exchanges in the past 30 days, three shared calendar events, and four mutual connections scores in the top quartile of relationship strength. At the other end, a contact with no interaction in 90 days and no shared connections scores near the bottom and should be treated as cold.

The signal everyone under-weights is reciprocity, and it is where scoring models lie to you.

Here is how each signal proves strength, and what it looks like when it lies:

• Recency proves the relationship is live. It lies when a high historical email count sits on top of zero contact in the last 90 days - a once-close tie that is now cold.
• Frequency proves ongoing contact. It lies when the volume is one-directional; see the reciprocity rule above.
• Shared connections / network depth proves the tie sits in a real cluster. It lies when the "connection" is a LinkedIn 1st-degree link with no interaction record behind it.
• Tenure overlap proves the two people actually worked together. It lies rarely, but confirm the overlap dates rather than assuming a shared employer means a shared desk.

In my example, the LinkedIn-only path scored high on network depth and near zero on everything else - a vanity path. One former colleague looked strong on frequency until I checked direction and found the replies had dried up eight months ago. The recent job-changer scored top-quartile on recency and reciprocity, with genuine tenure overlap on my team. That became the top of the shortlist.

This is the point where a tool earns its place. Enumerating who your whole team knows at one account, and scoring those ties by real interaction rather than LinkedIn degree, is slow by hand and is exactly what relationship-graph search is for.

Once you have a ranked shortlist, you verify the top tie before you spend anyone's reputation on it. Naming the second-degree buyer precisely and pooling the whole team's graph is where a tool like Refolk removes the manual grind, so the hours go into the ask, not the lookup.

The procedure, end to end

This is the full sequence for one account, from a blank logo to a booked meeting or a cold-queue decision. Follow it with your hands.

On sequencing, there is genuine disagreement worth knowing. Relationship-intelligence vendors say check the warm path before any cold outreach. Joanne Black of No More Cold Calling argues the referral ask should happen in a live conversation, not by email. Getro's workflow filters the account list first, then picks the connector. I use the order above because it moves from cheap, reversible work (naming and scoring) to expensive, one-shot work (spending a connector's capital) - but if your connector is a close colleague sitting nearby, take Black's advice and ask them out loud.

Drafting the ask and the forwardable blurb

Write two things: a short private note asking the connector's permission, and a forwardable blurb written to the target. The second one is the one that matters, because whatever you send to the connector will be forwarded to the target without edit. Write it as if the buyer is reading it, because they will.

The structure is fixed: subject line with both names, one line of context per person, one explicit reason they should talk, one clear next step. Keep the blurb near 120 words, then move yourself to BCC once the intro lands.

Subject: Quick favor - intro to [Buyer first name]?

Hi [Connector], hope you're well. I'm working on getting in front of [Buyer full name], [Buyer title] at [Target]. I noticed you two are connected - would you be open to a double opt-in intro? I've drafted a short forwardable blurb below so it's zero effort for you. Totally fine if the timing's wrong or you'd rather not. Thanks either way.

Hi [Buyer first name],

[Connector] suggested I reach out. I run [role] at [my company]; we help [one-line, concrete outcome for a company like theirs]. I follow [Target]'s work on [specific, real detail] and had one idea I think is worth 15 minutes of your time - no pitch deck, just a question about how you're handling [specific problem].

Would a short call in the next couple of weeks make sense? Happy to work around your calendar.

Thanks,
[Your name]

Two rules keep this from backfiring. First, keep the first ask small: a 15-minute chat, not a 90-minute architecture review. Big asks belong in the second email, after the intro. Second, send intros on Monday afternoons or Tuesday mornings - past the Monday inbox crush, with the full week left for a same-week reply.

Whatever you send the connector reaches the buyer unedited, so write the blurb to the buyer, not to your friend.

Running the double opt-in without burning the connector

A double opt-in introduction confirms interest from both parties before the intro is completed: the connector pings the requester and the target separately, gets a clear yes from each, then sends the connecting email. Run well, it holds intro acceptance at 60-80%. It is not bureaucracy; it is a reputation firewall.

The reason to hold this line is that if the target is annoyed by the intro, the connector takes the blame. Skip the double opt-in for speed and you burn the person who helped you - and connectors who get burned stop helping. The flow also protects everyone when the answer is no: if the target declines, they can do it off-thread, and the connector simply tells you it did not work out. No harm done, no awkward group email.

Follow up with the connector after about seven days of silence, once, gently. If the connector ghosts, that is a signal too - move to your second path.

How this goes wrong

The failure modes here are specific, and each has a false positive you can check for. This section is the most valuable part of the guide, because every one of these has cost a rep an account or a relationship.

• Volume mistaken for strength. A connector who sent 40 emails and got 2 replies looks strong in a naive CRM view. It is a weak relationship dressed as a strong one. Check reciprocity before ranking.
• Stale tie. A once-close contact with no contact in 90-plus days is effectively cold, no matter how high the historical count. Check the last-interaction date.
• LinkedIn-only connection. A 1st-degree LinkedIn link is not a relationship. Some people will not respond even when you were certain they would. Verify with a real interaction record.
• Connector declines or ghosts. Even with opt-ins all around, intros have a fairly high failure rate. Do not bet one account on one path; line up a second connector before you ask the first.
• Skipping the double opt-in for speed. Burns the connector, who takes the blame if the target is annoyed. Never blind-CC two busy people and hope.
• Asking for too much. Double opt-in works for a 15-minute chat, not a 90-minute review. Keep the first ask small; the big ask belongs in the follow-up.
• Wrong buyer named. A vague "anyone at Acme" stalls, and a mis-identified buyer wastes the connector's capital. Confirm the individual and their current role before you ask anyone.

When to stop looking and go cold

If no path clears the strength threshold, route the account to signal-based cold outreach rather than spending a connector's capital on a weak tie. A precise numeric cutoff is not established publicly, but the usable heuristic from scoring models is clear: a tie with no meaningful interaction in 90 days and no reciprocity scores at the bottom and should be treated as cold.

Do not treat "I found no path" as final on your own graph alone. An individual rep has no warm path to roughly 60-70% of accounts; pooled across the company, that drops to about 20-40%. So the honest sequence is: check your graph, check the team's, check for a recent job-changer, and only then declare the account cold. For a UK target, where the senior-buyer pool is about 17.8x smaller than the US, expect to hit the cold case more often and lean harder on pooling.

Going cold is not a failure state. It is a routing decision that saves your best connectors for the accounts where a real path exists.

What to verify before you send

Run this checklist before the intro leaves the connector's outbox.

Keeping the work current

A warm path is a perishable asset. The recent job-changer who is your best connector today ages out of their first-90-days window; the reciprocal tie that scored top-quartile this quarter goes quiet next quarter. Re-run the enumeration and scoring whenever an account moves into an active cycle, and re-check the recency date on any path older than 90 days before you use it.

The one input that keeps improving is the pooled network. Every new hire, customer, and partner adds nodes, and job changes constantly open new entry points into accounts that were cold last month. Set a standing check for people who join your target accounts and previously overlapped with your team - that single signal converts more cold logos to warm than any other, and it decays fastest if you do not watch it. Since willingness to refer runs at 83% while only 29% are ever asked, the accounts you lose are almost never lost for lack of a path. They are lost for lack of the ask.