Building a per-client channel-preference layer that stops LinkedIn when email is preferred

The prospect replied on email: "Email me, I barely open LinkedIn."

Two days later the sequence fired a connection request anyway.

That is not multichannel. That is ignoring a stated preference because the sequence template still had a LinkedIn step. Agencies running many client ICPs need a preference layer per workspace, not a global "always dual-channel" habit.

Preference is not the same as opt-out

Allston Labs' list-hygiene framing draws a useful line:

• Hard do-not-contact / statutory opt-out — stop everywhere, every channel
• Channel preference — stop or skip one channel, keep another if allowed
• Global unsubscribe click — policy choice: treat as all-channel or email-only

Outreach's granular opt-out docs show the product pattern: email, call, and text can be opted separately, and rulesets decide whether a channel opt-out finishes the sequence or only skips that channel's steps. Agencies need the same idea for LinkedIn vs email, even when the underlying tools differ.

SalesTarget and Flowkon both argue LinkedIn and email do different jobs. Cykel's multichannel piece is direct on the reply rule: if they answered on email, continue on email. Do not drag them onto LinkedIn because your cadence still has a connection step left.

Build the layer as fields, not tribal knowledge

In each client-isolated workspace, store preference at the person (or account) level:

unknown means the default sequence can still use both channels under the client's playbook. email_only must block LinkedIn connection requests, InMails, and profile-view automation for that person, not just "try LinkedIn softer."

Inferred behavior is weaker than a stated preference. Opens without replies are not consent to keep hammering LinkedIn. A written "email me" is.

Wire it into the sequence engine

Before every LinkedIn step:

1. Load dnc_all and channel_pref for that person in this client workspace.
2. If dnc_all or preference excludes LinkedIn, skip the step and keep email steps that remain allowed.
3. If preference is email_only after a positive email reply, also stop any pending LinkedIn tasks already queued.
4. Log the skip reason so QA can see preference enforcement, not a mysterious gap in the cadence.

Mirror the check before email steps when preference is linkedin_only (rarer, but real for some ICPs).

Keep suppression ledgers contact-keyed across systems, as Allston Labs argues. A LinkedIn-only preference that lives only inside one sequencer will reappear after the next enrichment refresh if the CRM never learned it.

Per-client policy, not one agency default

Client A may require dual-channel for enterprise ABM. Client B may forbid LinkedIn automation entirely. Client C may allow LinkedIn until the prospect states email preference, then lock.

That policy belongs in the workspace ruleset, next to send caps and voice briefs. Ops should be able to answer, for any client:

• What is the default channel mix?
• What happens on "email me"?
• What happens on a LinkedIn unsubscribe-style request or "please don't message here"?
• Does a hard email unsubscribe also kill LinkedIn for that client, or only email?

Write the answers down. Do not leave them in Slack folklore.

Failure modes

• Preference stored on the company record while two contacts disagree
• LinkedIn seat owner manually messaging someone marked email_only
• Preference updated in CRM but not synced before the next LinkedIn batch
• Treating "OOO, email me next week" as a permanent email_only lock without an expiry

GTM's motion runs email + LinkedIn inside client-isolated workspaces, stops sequences on reply, and keeps a compliance gate in front of send. Channel preference is how you respect the reply without pretending every prospect wants both channels forever.

If you want preference-aware outbound per client workspace, join the waitlist.