How to prevent intro collisions during a fundraise

The four collisions, and why each one costs you

Not every collision looks the same, and they fail in different ways. Naming them is the first step, because you cannot prevent a category you have not named.

The duplicate intro is the one founders notice, because it is visible. The other three are quieter and arguably worse. A stale intro means you let someone spend social capital on a door that is already closed or already open, which makes you look like you are not tracking your own round. The wrong-partner collision is expensive inside the fund: two partners now each have a partial picture, and the internal version of "who owns this deal" gets muddy before you even have a champion. The uncoordinated advisor is the most dangerous, because the investor's first real impression of your company comes from someone who is not you, framed in a way you never approved, and you find out only when the meeting starts cold and slightly off.

The common thread: every collision happens because two people both believed they had the right to act on the same investor. That belief is the bug. Helpers do not collide when they know who owns the path.

The operating principle: one owner, one active path

The rule is two sentences. Every investor on your list has exactly one owner, the person responsible for the current move toward that investor. Every investor has at most one active path open at a time, and a new path does not open until the current one is closed or explicitly handed off.

Owner does not mean the most senior or best-connected person. It means the person whose warm path you are using right now. If your advisor has the strongest relationship with a partner, the advisor owns that path. If a portfolio founder offered first and the relationship is good enough, that founder owns it. The owner is whoever is making the next move, and there is only one of them per investor at any moment.

"One active path at a time" is the rule that actually prevents duplicates. It is tempting to think two warm paths to the same partner double your odds. They do the opposite. A second simultaneous path does not add a second vote, it adds confusion the investor has to resolve, and the resolution is usually "I will wait until this settles down," which means you wait. Run the paths in sequence, not in parallel. Use the strongest one first. Keep the second in reserve in case the first goes quiet, at which point you hand off ownership cleanly and the second path opens.

The thing that makes this real instead of aspirational is that it has to be written down somewhere every helper can check before they act. A rule that lives only in your head is not a coordination system, because the whole problem is that the other people helping cannot see inside your head.

The intro path ledger

This is the artifact. One row per investor, owned and updated by you, visible to everyone helping you raise. Before anyone makes a move toward an investor, they check this. It is the single source of truth for who owns what and what is in flight.

INTRO PATH LEDGER

Investor: [Partner name] @ [Fund]
Status:   [Not started / Path open / In process / Passed / On hold]
Owner:    [Who is making the current move. Exactly one name]
Active path:   [The specific warm path being used right now]
Reserve path:  [Backup connector, used only if active path stalls]
Last touch:    [Date + what happened]
Next move:     [The one next action + who does it + by when]
History:       [Prior contact, prior pass, prior meeting, so nobody re-opens a closed door]
Do-not-intro:  [Yes/No. Set Yes when a path is already active, to block collisions]

The two fields that prevent most collisions are Owner and Do-not-intro. Owner answers "is this path already taken?" Do-not-intro is the flag a helper sees that tells them, in one word, not to act even if they have a great connection. The History field prevents stale intros: before anyone offers to open a door, they see that you already met this partner in March, or already passed in this fund, and they spend their capital somewhere useful instead.

Fill the History field even for investors you have not contacted yet, with the prior context you do have: "met at [event] last year," "passed on us at pre-seed," "advisor knows the other partner." That is the field that stops a helper from re-introducing you to a fund that already said no, which is the collision that makes you look least in control of your round.

The connector communication scripts

The ledger only works if you actually talk to the people offering to help, before and during their move. Most collisions happen in the silence between "I'd love to intro you to [partner]" and the email going out. Close that silence with three short scripts.

When someone offers an intro you want, and the path is open, claim it and set the terms:

"Yes, that one would be great. You'd be the owner of that path, so before you send anything, two things: I'll send you the blurb to forward so the framing is consistent, and please loop me before it goes out so I can confirm we don't already have something in motion there. Cool to wait for that?"

When someone offers an intro to an investor whose path is already active with someone else:

"Really appreciate it. We actually already have a path open to [partner] right now, so let's hold yours in reserve. If that one goes quiet in the next couple of weeks, I'll come back to you and you'd take over. Keeping it to one active intro per fund so we don't land in their inbox twice the same week."

When someone offers an intro to a fund you already passed with or are mid-process with:

"Good instinct, but heads up: we already [met / are in process / passed] with [fund] through [partner]. So an intro to [other partner] would actually collide with that. If it'd help, the more useful thing would be [a different fund / a reference call / an intro to X instead]."

The pattern across all three: thank them, tell them the truth about the current state, give them a specific alternative so the offered help gets redirected instead of wasted. Helpers do not get offended by "we already have that covered." They get offended when they spend their capital and later learn it collided with three other people doing the same thing.

The one rule that ties the scripts together: no helper sends an intro on your behalf without looping you first. Not because you do not trust them, but because you are the only person who can see the whole board. "Loop me before it goes out" is the single sentence that converts a crowd of well-meaning helpers into a coordinated round.

Where this gets expensive at scale

For five investors and two helpers, you can hold the ledger in your head and a shared doc. The coordination cost stays low because the board is small.

It breaks around the point where a founder-led round actually lives: thirty to sixty investors, several advisors, a few portfolio founders offering paths, intros arriving over email and DMs and hallway conversations across weeks. Now the question "is there already an active path to this partner?" has an answer that is scattered across your inbox, your advisor's inbox, three text threads, and a meeting note from two weeks ago. By the time you reconstruct it, the duplicate intro has already gone out. The ledger is the right idea, but keeping it current by hand across that many sources and people is exactly the work that slips when you are also taking meetings.

This is the part of the workflow RoundOS is built to carry. It reads the sources where your round already lives, your email, calendar, meeting notes, and investor lists, and maps the relationship graph: who has a path to which partner, which paths are already in motion, and which funds you have already touched or passed with. So when a new intro offer comes in, you can see in one place whether that path is already owned, already active, or already stale, before you say yes. The ledger stays yours to decide. What you stop doing is the cross-source reconstruction of "wait, did we already reach out there?" that lets collisions through.

The point is not to centralize for its own sake. It is that a collision is just two people acting on the same investor without a shared view, and a shared view is exactly what scatters across sources as the round grows.