Designing an Outreach Cadence That Gets Replies

Most people design their outreach cadence backwards. They write the messages first, then figure out the timing, then wonder why replies drop off after the first follow-up. The better order: decide your branch logic before you write a single word, because the structure determines whether a sequence feels like a conversation or a harassment campaign.

Here is the playbook we use and refine constantly.

What an Outreach Cadence Actually Is

A cadence is not a drip sequence. A drip sequence blasts everyone the same messages on the same schedule. A cadence is conditional: it responds to what the prospect does, or does not do, at each step.

The practical difference matters. If someone accepts your connection request and immediately views your profile twice, treating them identically to someone who has not logged into LinkedIn in a week is a wasted opportunity. Branch logic fixes that.

A solid LinkedIn outreach cadence has three components working together: the steps (what you send), the delays (when you send it), and the branches (what triggers a different path). Nail all three and the sequence does the judgment calls without you staring at a spreadsheet every morning.

Before building anything, make sure your profile is set up for the messages you are about to send. A polished first message lands differently when the profile behind it looks credible. Optimizing Your Profile for Outbound: A Playbook covers that in detail.

The Steps: What to Send and When

Five touchpoints is our baseline recommendation for cold outreach. Seven if the deal size justifies more patience. Beyond seven, you are likely annoying more people than you are converting.

Step 1: Connection request with a short note (Day 0) Keep it under 200 characters. One specific reason you are reaching out. No pitch. The mistake we keep seeing is founders treating the connection note as the pitch itself, which is the fastest way to get ignored or declined.

Step 2: First message after acceptance (Day 2-3) This is your actual opener. One paragraph, one question, no attachment. Send it 2-3 days after they accept, not the same hour. Immediate follow-up feels automated even when it is not.

Step 3: Value-add follow-up (Day 6-7) If no reply, send something useful: a relevant stat, a case study, a short observation about their industry. Not another version of your pitch. This step exists to give them a reason to reply that is not "yes, I want to buy your thing."

Step 4: Soft bump (Day 11-12) Short. Three sentences maximum. Something like: "Wanted to make sure this did not get buried." No new content, no pressure. Just visibility.

Step 5: Breakup message (Day 18-21) The breakup message has the highest reply rate of any step in our sequences. Be direct: "I will take your silence as a no for now and will not message again. If timing changes, my door is open." People reply to breakup messages because it feels like a closing door.

Delays and Timing: The Numbers That Actually Matter

Timing is where most cadences go wrong. People either send too fast (looks automated, feels pushy) or space things out so long the prospect forgets who you are.

We cap our own sends at 20-25 connection requests per day. That is not a conservative safety buffer: it is the actual ceiling where LinkedIn's anomaly detection stays quiet in our experience. Go higher and you will see connection acceptance rates drop within a week as the algorithm deprioritises your invites.

Randomised timing jitter matters too. Sending every message at 9:00 AM sharp is a machine tell. Spread sends across a 2-3 hour window around your target time and vary which days you are active.

Branch Logic: Where the Intelligence Lives

This is the part most tools ignore or bury behind advanced settings. Branch logic is what separates a sequence that adapts from one that just fires.

The core If/Else decisions in any LinkedIn cadence:

If they reply at any step: Stop the sequence immediately. Auto-pause is non-negotiable. Continuing to send automated messages after someone has responded is the fastest way to kill a deal and, frankly, it is just rude.

If they accept but do not reply within 24 hours: Continue to Step 2 on schedule.

If they view your profile after Step 3 but have not replied: Branch to a shorter, more direct message. Profile views after a follow-up usually signal interest without enough friction-reduction. They need one more specific reason to respond.

If they do not accept the connection request within 7 days: You have two reasonable options. Send a follow-up InMail if you have credits, or withdraw and re-approach in 30 days with a different angle. Do not just leave unaccepted requests piling up.

If they click a link you shared: Flag them for priority manual follow-up. Link clicks are the strongest intent signal in a cold cadence.

For a deeper look at the full booking workflow these branches feed into, How to Book Meetings from LinkedIn: A Playbook walks through the conversion layer after the reply comes in.

Balancing Persistence and Politeness

There is a real tension here and pretending it does not exist is not useful. You need enough touchpoints to catch people on a good day, but too many and you are training prospects to tune out your name.

Our rule: every message has to earn its place. If you cannot articulate why Step 4 exists independently of Step 3, cut Step 4. The cadence should never feel like you are waiting for someone to cave.

Personalisation helps, but it is not a replacement for structural discipline. A well-timed generic message beats a poorly timed personalised one. Get the structure right first, then layer in personalisation at the message level.

One thing that genuinely shifts the balance: A/B testing your openers. In our own testing, changing the first sentence of the connection note moves acceptance rates noticeably. Test one variable at a time. Opening line one week, the value proposition framing the next. Never change two things simultaneously if you want clean data.

How Ampliflow Handles This in Practice

This is the one section where the tool is directly relevant to executing everything above.

Ampliflow's workflow builder is visual and drag-and-drop. You lay out each step as a node, set the delay on each connection, and drop in If/Else branches wherever you need conditional logic. The breakup-message branch, the profile-view trigger, the reply-detection auto-pause: all configurable without writing code or stitching together a Zapier chain.

Cloud execution matters for cadence design specifically because timing jitter only works if your sequence runs even when your laptop is closed. Browser extensions cannot do that reliably. Ampliflow runs via the Unipile API, so messages fire at scheduled randomised times whether or not you are online.

The built-in account safety scoring shows a real-time risk indicator based on your daily activity. If you have been more active than usual on a given day, the dashboard flags it before you breach the threshold, not after. That feedback loop changes how you think about cadence pacing.

Pricing starts at $19 per month for founding members, with the first 100 locking that rate for life. Public launch pricing moves to $39/month for Starter and $79/month for Pro. For context, Dripify starts at $79/month and Expandi at $99/month. Cheaper tools like Linked Helper ($15/month) and Dux-Soup ($14.99/month) are genuinely cheaper, and it is worth being honest about that. The trade-off is architecture: those tools run in your browser, which means a different risk profile and no persistent cloud execution. Worth knowing before you choose.

If you are already evaluating alternatives, Dripify Alternative: Cloud LinkedIn Automation From $19/mo breaks down that comparison in detail.

Common Mistakes Worth Naming

Pitching in the connection note. It signals impatience and tanks acceptance rates before the conversation starts.

Sending follow-ups without adding anything new. "Just checking in" is not a follow-up. It is proof you ran out of ideas.

No auto-pause on reply. We have seen prospects continue to receive automated messages after they responded expressing interest, and then they went cold. That is fixable and inexcusable.

Treating every prospect identically regardless of behaviour. If someone has viewed your profile three times and clicked your link, they do not need the same next message as someone who has shown zero engagement. Branch logic exists for exactly this reason.

Ignoring funnel drop-off by step. If Step 3 has poor open rates, the problem is either timing or the message before it. Check your funnel analytics after the first 50 sequences complete. You cannot fix what you are not measuring.

Designing an outreach cadence is fundamentally a systems problem, not a copywriting problem. Write decent messages, yes. But get the timing, the branches, and the daily limits right first, and the sequence will perform even on days when your copy is only pretty good.