Email and LinkedIn outreach works best when both channels run from one prospect record, one reason for reaching out, and one sequence that stops the moment someone replies. The email carries the argument, LinkedIn makes the sender a real person, and a phone step (when it fits) shortens the path to a conversation. This guide shows how to set that up, with sending windows, stop-on-reply rules and a sample cadence you can adapt.
Nerdy Joe is owned by Agency Growth Partners (AGP), so this is our own system and we are not neutral about it. This post is one stage of the AGP growth machine, the end-to-end outbound system AGP runs for marketing agencies on its Artemis platform. The examples below describe how Artemis handles multichannel, but the rules apply to any tool or DIY build.
Why email and LinkedIn outreach belong in one sequence
Most agencies that try both channels run them as separate projects. Someone sends cold email from one tool, someone else sends LinkedIn connection requests from another, and neither knows what the other said.
The prospect notices. They get an email about a hiring push on Monday, a generic LinkedIn note on Tuesday that ignores the email, and a follow-up email on Thursday after they already replied on LinkedIn. That reads like two vendors, not one thoughtful sender.
Running the channels in sync fixes three things:
- One story. Every touch builds on the last one instead of starting over.
- One stop. A reply anywhere ends the automated steps everywhere.
- One record. The full conversation history sits on the prospect, which matters later when someone has to take the meeting.
Start from the same signals on every channel
The reason you reach out should be the same on email and LinkedIn. If the email references a new plant opening, the LinkedIn message should not pivot to a generic "would love to connect."
Inside Artemis, messages are written from live signals such as role, company stage, hiring activity, a new plant opening or a hiring push, and then adapted for email or LinkedIn. LinkedIn messages are personalized from the same live signals as the email and queued alongside it. The previous stage, signal-based cold email personalization, covers how those signals are found and turned into copy.
What changes between channels is format, not substance:
| Element | ||
|---|---|---|
| Length | Short, but room for a full point and a clear ask | Much shorter, often one or two sentences |
| Tone | Direct and specific | Conversational, closer to a note than a pitch |
| Signal use | State the signal and why it matters to them | Reference the signal lightly, assume they may have seen the email |
| Ask | A specific question or a meeting | Usually a connection, a reaction or a soft question |
| Links | Sparing, often none in the first touch | Usually none |
A useful test: if you deleted the prospect's name and company, would the LinkedIn message still make sense? If yes, it is not using the signal and it will read like spam.
Where phone fits in a multichannel sequence
AGP's services page lists phone alongside email and LinkedIn in its outbound sequences. Phone is the channel most agencies skip, and also the one that can turn a warm-but-silent prospect into a conversation.
Phone earns its place when:
- The deal size is big enough to justify a person's time on each call.
- Your buyers actually pick up (owners and operators often do; some enterprise titles rarely do).
- The caller can see the email and LinkedIn history before dialing, so the call continues a conversation instead of starting a cold one.
Phone does not belong in the sequence when nobody is assigned to make the calls. A call task that sits untouched for two weeks breaks the cadence timing and clutters the pipeline. Either staff it or leave it out.
Multi-step sequences: how the steps fit together
A multichannel sequence is a set of timed steps for each prospect. Each step has a channel, a delay after the previous step, and a condition that decides whether it runs. Inside Artemis these are multi-step sequences with follow-ups, stop-on-reply and sending windows, and email and LinkedIn are kept in sync.
Build it in this order:
- Write the core argument once. One sentence on why you are reaching out to this person now, based on the signal.
- Map it to channels. Decide which channel carries the full argument (usually email), which one adds familiarity (usually LinkedIn), and whether phone is in play.
- Set the delays. Leave enough space that each touch feels separate. Back-to-back touches on the same day across channels feel automated.
- Add conditions. For example: only send the LinkedIn message if the connection request was accepted; only queue the call if the prospect has a phone number on record.
- Write follow-ups that add something. A new angle, a short example, a different question. "Just bumping this" is not a step.
- Decide the exit. After the last step, the prospect leaves the active sequence and moves to a longer follow-up track or is closed out.
Sample multichannel cadence (illustrative)
This is an illustrative example of a three-week cadence, not AGP's standard sequence or a benchmark. Adjust the days and channels to your market.
| Day | Channel | Step | Purpose |
|---|---|---|---|
| 1 | First email, built on the signal | Make the core point and ask one question | |
| 2 | Profile view, then connection request with no pitch | Put a face to the name | |
| 4 | Follow-up with a different angle | Add a concrete example or outcome type | |
| 6 | Short message if connected, referencing the same signal | Low-pressure second channel | |
| 8 | Phone | Call, with the email and LinkedIn history on screen | Try a live conversation (only if staffed) |
| 11 | Short follow-up with a direct question | Make replying easy | |
| 15 | Light touch, such as a comment or a one-line note | Stay familiar without pushing | |
| 19 | Closing email that states you will stop | Give a clear, polite exit |
Stop-on-reply applies to every row. If the prospect answers on day 4, days 6 through 19 never happen for them.
Stop-on-reply: the rule that protects your reputation
Stop-on-reply is the single most important setting in a multichannel sequence. When a prospect replies on any channel, every remaining automated step for that prospect should pause.
Without it, three bad things happen:
- A prospect who said "not interested" gets two more emails and a LinkedIn message, and some of them mark you as spam.
- A prospect who said "yes, send times" gets a "just following up" email the next day, which makes the sender look careless.
- Your reply data gets messy because you cannot tell which touch actually earned the response.
Check these details in any tool you use:
- [ ] Does a LinkedIn reply stop the email steps, and does an email reply stop the LinkedIn steps?
- [ ] Does an out-of-office reply pause the sequence or restart it later? (Treat it differently from a real reply.)
- [ ] Does a reply from a colleague at the same company stop the sequence for others there?
- [ ] Does an unsubscribe on one channel suppress the person everywhere?
The next stage, AI reply handling and follow-up, covers what happens to the reply once the sequence stops.
Sending windows and spacing
Sending windows control when steps go out. Spacing controls how far apart sends are within that window.
Inside Artemis, sends are spread across mailboxes under daily limits, with sending windows and spacing between sends. Volume drops automatically when a mailbox starts to slip into spam. That protects deliverability, which is covered in detail in mailbox warmup and inbox placement.
Practical rules for setting windows:
- Use the prospect's time zone, not yours. An email that lands at 3 a.m. local time sits under a pile of newer messages by morning.
- Send during business hours on business days for most B2B buyers, and test narrower windows once you have reply data.
- Space sends out. A mailbox that fires 40 emails in two minutes looks nothing like a person. Spread them across the window.
- Keep LinkedIn activity human-paced. Bursts of connection requests are the pattern platforms watch for.
A note on LinkedIn limits
You will see many specific numbers quoted online for how many LinkedIn connection requests or messages you can send per day or week. LinkedIn does not publish a simple fixed number we can cite, and the figures in circulation change and vary by account. We are not going to repeat them here.
What we can say: LinkedIn sets its own rules on automated activity, so read its current terms before you automate anything. Keep activity at a pace a real person could plausibly keep, warm up new accounts slowly, and walk away from any provider that promises unlimited LinkedIn volume.
Midpoint check: where multichannel sits in the machine
Multichannel outreach only works if the stages before it are solid. A sequence across three channels will not save a bad list, an unclear offer or mailboxes landing in spam. That is why how AGP's growth machine works puts strategy, data, lists, sending infrastructure and warmup ahead of this stage, and reply handling and booking right after it.
Here is the order, with this stage in context:
| Stage | What it supplies to multichannel |
|---|---|
| Ideal client profile and offer | Who to contact and why they should care |
| Data and verified lists | Accurate emails, phone numbers and LinkedIn URLs on one record |
| Sending infrastructure and warmup | Mailboxes that land in the inbox |
| Personalization | The signal-based message each channel adapts |
| Multichannel outreach | Timed touches across email, LinkedIn and phone |
| Replies and follow-up | What happens when someone answers |
| Booking and prep briefs | Turning interest into a prepared meeting |
Testing and optimizing a multichannel sequence
Inside Artemis, campaigns self-optimize toward messages that earn positive replies, with A/B variants per step. AI can draft and rewrite copy (shorten it, make it friendlier, fix grammar) and score email content. An Autopilot mode can hold emails for human review before they send, which is useful in the first weeks of a new campaign.
Whatever tool you use, test one thing at a time:
- The first email first. It carries the most weight. Test two angles on the same signal before you touch later steps.
- Then channel order. Email first versus LinkedIn first, on comparable slices of the list.
- Then the follow-ups. Which second angle earns more replies.
- Then timing. Gaps between steps and sending windows.
Judge tests on positive replies, not opens. A step that gets many replies saying "remove me" is not winning.
Common mistakes in email and LinkedIn outreach
- Two tools, no shared record. The prospect gets contradictory messages and nobody can see the full thread.
- The LinkedIn note is a pasted email. Long pitches in a connection request rarely land.
- No stop-on-reply across channels. The fastest way to annoy someone who already answered.
- Phone steps with no caller. Tasks pile up and timing breaks.
- Same-day touches on every channel. It feels like being surrounded, not contacted.
- Ignoring time zones. Sending at your local 9 a.m. to a list spread across several time zones.
- Measuring opens. Judge every channel on replies and booked meetings.
How to judge a provider or DIY build
Use these questions on AGP, any other agency, or your own setup:
- [ ] Do email, LinkedIn and phone steps live in one sequence on one prospect record?
- [ ] Is every message on every channel built from the same signal?
- [ ] Does a reply on any channel stop all remaining steps?
- [ ] Are sending windows set to the prospect's time zone, with spacing between sends?
- [ ] Is LinkedIn activity paced like a person, with no promise of unlimited volume?
- [ ] Is a named person assigned to phone steps, or are they left out?
- [ ] Can you see the full multichannel history for any prospect at any time?
- [ ] Are tests judged on positive replies and meetings?
If you would rather have LinkedIn handled as its own service, our page on LinkedIn lead generation services explains that option. For the email side of the build, see our guide to cold email infrastructure.
Who this approach is not for
Multichannel outreach adds coordination. It is not the right move for everyone:
- Very small lists. If you have 50 target accounts, hand-written, one-to-one outreach by the founder may beat any sequence.
- Teams without LinkedIn profiles worth visiting. If the sender's profile is empty, the LinkedIn touch can hurt more than help.
- Offers that are not clear yet. More channels repeat an unclear message more often. Fix the offer first.
- Anyone who needs meetings this week. Mailboxes and accounts need time to warm up before volume is safe.
Want this built for your agency?
AGP builds and runs this stage, along with the rest of the full agency growth machine, on Artemis for marketing agencies. A dedicated growth engineer builds, runs and tunes the sequences, copy and deliverability. If you want to see how it would map to your market, book a call with David.
Sources
- Agency Growth Partners, home: https://agencygrowth.partners/
- Agency Growth Partners, Artemis: https://agencygrowth.partners/artemis
- Agency Growth Partners, Services: https://agencygrowth.partners/services
Frequently asked questions
Should an agency run email and LinkedIn outreach at the same time?
Yes, if both channels are driven by the same prospect record and the same reason for reaching out. Running them as two separate campaigns usually means repeated messages, mixed signals and replies nobody connects back to the email thread.
Which should come first, the email or the LinkedIn touch?
Most multichannel cadences open with email because it carries the full message, then use a LinkedIn profile view or connection request a day or two later so the name is familiar. There is no single right order; test it on your own list.
How many steps should a multichannel outreach sequence have?
Enough to make the point a few different ways without nagging. There is no correct number; start with a handful of touches spread over a few weeks across email, LinkedIn and phone, test it, then stop and move the prospect to a longer follow-up track.
What does stop-on-reply mean in a sequence?
Stop-on-reply means the moment a prospect answers on any channel, every remaining automated step for that person pauses. A human or an AI step then handles the reply, so nobody gets a scheduled follow-up after they already said yes or no.
Is it safe to automate LinkedIn outreach?
LinkedIn sets its own rules on automated activity and enforces them on accounts. Read its current terms, keep activity at a pace a real person could keep, and treat any provider who promises unlimited LinkedIn volume as a risk.
Do I need phone calls in a B2B outbound sequence?
Not always. Phone tends to earn its place when deal sizes are large, buyers are reachable by phone, and someone trained can make the calls. If nobody will actually dial, leave it out instead of adding call tasks that pile up.