Running several LinkedIn accounts without two messages hitting one prospect

What overlap looks like, and why it costs more than it seems

It looks like two connection requests from two different strangers at the same vendor arriving in one week. To the prospect, that is not persistence, it is a machine. The first message might have been read as a person reaching out. The second one retroactively reclassifies both as a campaign, and the credibility you were buying by using a real profile with a real history is spent for nothing.

The internal cost is quieter and lasts longer. Two operators now have half a conversation each, the CRM has two records for one company, and whichever one closes will be attributed to the wrong campaign. Every number you report afterwards is slightly wrong in a way nobody can trace.

• A prospect replies "someone from your team already messaged me" and neither operator knows who.
• Two open threads sit in two inboxes with no shared view between them.
• The CRM holds duplicate records, so suppression checks pass when they should fail.
• Reply attribution lands on whichever record was updated last, not the one that did the work.
• A company you were carefully multithreading gets contacted by a third profile that knew nothing about it.

The fix is not asking people to be more careful. Careful people collide too, because nothing in the sending workflow tells them the collision is about to happen.

Route by rule, not by capacity

Assign prospects by a written rule that produces the same answer every time, whoever applies it. The failure pattern is assigning to "whichever account has room today", because then the assignment depends on timing, and nobody can reconstruct six weeks later why a company went to one profile rather than another — which means nobody can tell whether it already went to two.

Whichever you choose, write it as one sentence that a new operator can apply on their first morning without asking anyone. "Profile A takes healthcare and education, profile B takes everything else" is a rule. "Split it evenly" is not, because two people will split it differently. If you are still deciding how many profiles to run in the first place, that is a separate question covered in how many LinkedIn accounts to rent for outreach; this post assumes the count is settled.

Exclusion is account-level, not person-level

Deduplicate on the company, not the individual. Almost every tool deduplicates by profile URL, which happily allows one profile to invite the VP of Sales while another invites the CRO at the same company on the same afternoon. From inside the target company, that is the same collision — arguably a worse one, because the two of them talk.

• One company is in play with one profile at a time. That is the default rule.
• A company stays assigned to that profile for a defined window after the last touch, and the window is longer than your full follow-up sequence.
• Deliberate multithreading is allowed, but it comes from the same profile. Reaching three stakeholders from one persona reads as a salesperson doing their job. Reaching three from three personas reads as a list.
• An exception list exists for large enterprises where separate business units are genuinely separate buyers, and it is an explicit list, not an assumption.
• Company matching happens on domain, not on the display name, so "Acme" and "Acme Group Ltd" do not both get through.

If you run rented profiles, this rule is doing double duty. Besides protecting the prospect experience, it keeps each profile's activity looking like one person working a patch, rather than several personas converging on the same targets in a pattern no individual would ever produce. That is part of what a managed account rental arrangement is supposed to enforce on your behalf.

One master suppression list every account reads before it sends

One list, one owner, checked at two moments: when the list is built and again on the morning of sending. Deals close and complaints arrive in the gap between those two, and a list built on Monday is already slightly wrong by Thursday.

What belongs in it:

• Current customers, including anyone mid-onboarding.
• Open opportunities from any source, including inbound — the prospect does not care which channel found them first.
• Anyone who has asked not to be contacted, ever, with no expiry.
• Companies in a partner, supplier or reseller relationship with you.
• Anyone touched by any of your profiles inside your chosen re-contact window.
• Your own employees, your investors, and candidates in your recruiting pipeline if you also use LinkedIn for hiring.
• Competitors, if you have decided to exclude them, recorded as a decision rather than left to each operator's judgement.

The list has to live outside any single sending tool. Tools deduplicate within themselves, which is exactly the boundary the problem crosses: two profiles running in two tool instances will both pass their own checks and still collide. In practice that means the CRM is the source of truth and the sending tools read from it, not the other way round. It is the same reason the sending-profile field matters when you track LinkedIn-sourced pipeline at all.

When a prospect replies to profile B after ignoring profile A

The conversation belongs to the profile that got the reply, without exception. Do not reintroduce the first persona, do not hand the thread over "because that rep owns the territory", and do not have both profiles active in the same relationship. The prospect chose a person; changing the person now is the thing that breaks the illusion of a normal conversation, because normal conversations do not get transferred between strangers.

1. Profile B continues the conversation and owns it from that moment.
2. Profile A stops immediately — pending follow-ups cancelled, and any unaccepted invite withdrawn.
3. The CRM record updates the sending-profile field to B, with a note recording that A touched it first and on what date.
4. The company goes onto suppression for every other profile for the duration of the conversation and the window after it.
5. If the prospect asks whether someone else contacted them, answer plainly in one line and move on.

That last point is worth getting right in advance, because operators improvise badly under pressure. Something like: "Yes, we are both at the same company and our lists overlapped. I will be your only point of contact from here." It is true, it is short, and it does not make the prospect feel handled. Denying it, or pretending the two profiles are unrelated, is how a small routing error turns into a trust problem.

The weekly ritual that keeps it working

1. Monday: build the week's lists centrally, apply suppression, and assign by the written rule. Nobody builds their own list.
2. Monday: lock assignments for the week. A company assigned to a profile stays there even if that profile is slow.
3. Midweek: run one duplicate check across all profiles on company domain, not on person.
4. Friday: return untouched prospects to the pool so they can be reassigned next week under the same rule.
5. Monthly: review collisions that happened anyway, and fix the rule rather than reminding people to concentrate.

One more constraint if the profiles are rented rather than owned. LinkedIn's User Agreement asks members not to share their accounts or let others use them, and accounts can be restricted, so a profile can leave your rotation at short notice. Your routing rule needs an answer for that: which profile inherits the segment, whether the companies already touched go on hold, and how a replacement persona is introduced to live conversations. Agreeing that before it happens is far easier than improvising it during a campaign — talk to us if you want the rotation designed rather than assembled.

Key takeaways

• Assign prospects by a written rule a new operator could apply, never by who has capacity.
• Deduplicate on company domain, not on the individual person.
• Keep one master suppression list outside any single sending tool, checked at build and at send.
• Whoever gets the reply owns the conversation; the other profile stops that day.
• Plan in advance which profile inherits a segment if an account has to be replaced.

Frequently asked questions

Is it really a problem if two people at the same company hear from us?

It depends on who sent it. Two messages from one salesperson to two stakeholders is normal multithreading. Two messages from two unconnected strangers at the same vendor in the same week reads as automated prospecting, and it usually ends both conversations at once.

Can a tool handle deduplication for me?

Only inside itself. Most sending tools deduplicate against their own history, which is the wrong boundary when you are running several accounts, possibly in separate tool instances. Keep the master suppression list in your CRM and have every list built from there.

How long should a company stay assigned to one profile?

Longer than your full follow-up sequence, so a prospect cannot be reassigned while touch three is still pending. Pick a window, write it into the rule, and apply it to every profile rather than case by case.

What if two of our reps share one profile?

Avoid it. Beyond the coordination problem, you lose the audit trail of who sent what, and the account's activity pattern starts to look like two different people with two different schedules. One operator per profile is the simpler and safer arrangement.

We already collided with a prospect. What now?

Stop all pending activity from the second profile, acknowledge it in one sentence if the prospect raises it, and let the profile that received the reply continue alone. Then fix the routing rule, because the same gap will produce the same collision next month.