Email and LinkedIn sequence structure: which channel opens, and what follows
By Jānis Plūme, Founder, Outbound Pros · 2026-08-06
Quick answer
In a multichannel sequence, the opening channel should be the one with enough capacity to touch your whole list inside your sequence window, and the second channel should escalate onto the prospects who showed a signal. For most B2B lists above a few thousand names that means email opens and LinkedIn escalates, because email capacity scales with the number of mailboxes you run while LinkedIn capacity is capped per seat per day. The calculation is your prospect count divided by your daily LinkedIn ceiling. If that exceeds your sequence window in days, LinkedIn cannot be touch one.
What is a multichannel sequence structure?
A multichannel sequence structure is the assignment of channels to positions in a touch schedule: which channel makes first contact, which channel carries the follow ups, and which channel is reserved for prospects who have already done something. It is separate from copy and separate from targeting. Two teams can run identical lists and identical messages and get different results purely from which channel opened.
Structure gets decided badly in most teams because it gets decided by taste. Someone believes LinkedIn is warmer, so LinkedIn goes first. Someone else believes email scales, so email goes first. Both are describing a property of the channel and neither is describing the list.
Which channel should open a cold sequence?
The channel that can reach every prospect on the list inside the window should open. That is nearly always email, because email throughput is a function of how many mailboxes you have provisioned and LinkedIn throughput is a hard per seat daily ceiling that does not move when you buy more software.
Work it out before you design anything. LinkedIn coverage window in days equals prospect count divided by the daily connection request ceiling per seat, divided again by the number of seats.
Run that number against your intended sequence window. A 12,000 prospect list against a single seat is a coverage window measured in months. A sequence designed to run in three weeks cannot open on a channel that takes months to make first contact, so LinkedIn moves to the escalation position and email opens.
The ceilings themselves are a moving target and a research subject in their own right. Our sibling property maintains that research and a calculator for it at the LinkedIn safe sending calculator. We take whatever number comes out of it as a fixed input and plan the calendar around it.
Why is the better performing channel usually the second one?
Because response rate and capacity move in opposite directions, and the channel you have least of is the one that converts best per touch.
On one client programme in professional services, we measured 59% connection request acceptance, roughly 9% reply rate on LinkedIn DMs, and roughly 1.5% email reply rate on the same accounts over the same period. That dataset belongs to our sibling site, which publishes it in full with its own methodology note at linkedpros.io on DM reply rates.
The structural conclusion is the part that matters here. LinkedIn DMs replied at roughly six times the rate of email on identical accounts, and LinkedIn capacity is a small fraction of email capacity, because email throughput scales with mailboxes and LinkedIn throughput is a fixed per seat ceiling. A channel that returns six times more per touch and places far fewer touches is a precision instrument and not an opener, and pointing it at everyone spends it on the large majority of the list that was never going to engage. Structure exists to aim it.
That is the whole argument for sequencing instead of stacking. You spend the cheap, abundant channel discovering who is warm, and you spend the expensive, scarce channel on them.
What does a WideNET structure look like?
WideNET is our high volume motion, systematic angle testing across the full addressable market. Email opens, LinkedIn escalates on signal.
| Position | Channel | Fires when | Personalization budget |
|---|---|---|---|
| Touch 1 | Day 0, whole list | Segment level, not individual | |
| Touch 2 | Short gap after touch 1, no reply | None, same thread | |
| Touch 3 | LinkedIn connection request | Only for prospects who opened twice, clicked, or replied without a decision | High, this is the escalation |
| Touch 4 | LinkedIn DM | A short delay after acceptance, never immediately | High |
| Touch 5 | New thread, new angle, longer gap | Medium |
Note that touch 3 is conditional. The LinkedIn budget is spent only on prospects who did something. On a list where 5% show any engagement signal, a single seat can comfortably serve a list twenty times larger than its raw ceiling implies, because it never touches the 95%.
What does a Spearhead structure look like?
Spearhead is our signal triggered motion, running on the hottest slice instead of the full market. Here LinkedIn opens, because the list is small enough that it can.
| Position | Channel | Fires when | Personalization budget |
|---|---|---|---|
| Touch 1 | LinkedIn profile view | Day 0 | None, it is a visibility action |
| Touch 2 | LinkedIn connection request | Day 0 or day 1 | High |
| Touch 3 | Whether or not the request was accepted | High, references the signal that triggered the campaign | |
| Touch 4 | LinkedIn DM | Only if accepted, after a deliberate gap | High |
| Touch 5 | New angle, new thread | Medium |
The reason touch 3 fires regardless of acceptance is that the profile view and the pending request have already done work. The name is faintly familiar by the time the email lands, and the email is what carries the actual reason for contact. Prospects who accept get a fourth touch on the channel where they already said yes.
How long after a connection is accepted should you wait before the first DM?
Leave a deliberate gap rather than firing on acceptance. An instant DM is the clearest possible signal that the connection request was automated, because no human accepts a request and composes a message in the same second. Our default is a gap long enough that the message reads as a separate decision, typically the next working day, not the next minute.
The open web contradicts itself badly on this, mostly because vendors whose tools fire on the acceptance webhook have an obvious interest in the answer being immediately. What the message says once it fires is a copy question and belongs to our sibling site, which covers what a LinkedIn DM should say. When it fires relative to everything else is ours.
Where does this structure break?
When the two channels have different sender identities. If the emails come from one person and the LinkedIn touches come from another, you do not have a sequence, you have two vendors calling. Coordinate the identity or drop the second channel.
When the signal that triggers the escalation is noise. Open tracking is unreliable enough that escalating on opens alone puts your LinkedIn budget on people who never saw the email. We weight replies and clicks well above opens, and on accounts where tracking looks unreliable we escalate on reply only and accept a smaller escalation pool.
When the list is homogeneous. The whole structure assumes some prospects are warmer than others. On a list where everyone is equally cold and equally uninterested, escalating on signal escalates on randomness. That is usually a targeting problem wearing a cadence problem's clothes.
When there is no reply sync. A structure with two channels and no shared stop condition is worse than a single channel campaign, because the failure mode is visible to the prospect. Wire the cross channel stop first.
Get the structure without designing it
The Sequence Cadence Builder takes your prospect count, your seats and your window and returns the structure above with real dates on it. Free, no signup.
Both structures above are presets. The tool picks the entry channel from your own arithmetic rather than from a preference.
If you would rather hand the whole motion over, the group's agency runs both channels as one coordinated sequence, and the LinkedIn side of it specifically is the group's managed LinkedIn programme. Everything on this site is written from campaigns we run there, and the wider service picture sits at Outbound Pros.
Frequently asked questions
Should I send a connection request to everyone on my email list?
No, unless the list is small enough that your seat capacity can serve it inside your window. On lists larger than that, sending requests indiscriminately spends your scarcest channel on prospects who have shown nothing, and it drives acceptance rate down, which our sibling site covers as a mechanic in its own right.
Can I run both channels from day zero?
You can, and on small Spearhead lists we do. On large lists it produces a same day double approach for prospects who have never heard of you, which is the version of multichannel that reads as a blitz. The pairing works when it is deliberate and targeted, not when it is the default for everyone.
Does the email need to mention the LinkedIn touch?
No, and referencing it usually hurts. Saying that you also sent a connection request makes the sequence legible as a sequence. Let the touches accumulate familiarity without narrating the mechanism.
What if I have no LinkedIn seats at all?
Then you have an email sequence, and it should be designed as one, not as a multichannel sequence with a hole in it. The sending infrastructure, domain and warm up work that makes an email only motion viable is the parent agency's deepest specialty, and it is the part of the stack this site takes as given rather than teaches.
Last updated: 2026-08-06