How to Set Up a LinkedIn Outreach Sequence (2026)

Decision 1: Sequence Shape — Which Steps and in What Order

A LinkedIn outreach sequence is built from a small set of step types. The standard step types in cloud-based automation tools (HeyReach, Lemlist, Expandi, La Growth Machine, Skylead, Dripify):

• View profile — the account opens the prospect's profile. Counts as a soft touch — the prospect often gets a 'someone viewed your profile' notification, which warms them up before the connection request lands.
• Send connection request — with or without a personalized note. The core conversion step.
• Send message — only available after the connection request is accepted. Where the actual outreach happens.
• InMail — Sales Navigator only. Direct message to a non-connection. Separate channel from connection requests with its own monthly allotment.
• Follow up — a delayed message in the same conversation thread. Most replies happen on the second or third message, not the first.
• Conditional branches — if-accepted vs if-not-accepted paths, if-replied vs if-not-replied paths. How the sequence handles different prospect responses.

A Standard Sequence Structure

For most B2B outreach, the working sequence shape is:

Step 1: View profile. Soft touch, no risk. Day 0.

Step 2: Send connection request with personalized note. Day 1-2. Wait 3-5 days for acceptance.

Step 3: If accepted, send introductory message. Day 5-7 from connection. Short — thank them for connecting, mention something specific about their work or company, no pitch yet.

Step 4: Follow up with value message. Day 9-12. This is where you actually pitch — a specific value proposition tied to what their company does, ideally with a question that invites a reply rather than a CTA that demands action.

Step 5: Final follow-up. Day 16-21. Soft close — 'happy to share more if useful, otherwise no worries.' The message that converts highest is the lowest-pressure one because it filters for prospects who actually want to engage.

Sequences with more steps than this rarely outperform 5-step sequences. The conversion rate on step 6+ messages drops below 1% on most B2B campaigns, and the additional steps cost daily-volume capacity that's better spent on new prospects. For the multi-account scaling math that determines how many prospects this sequence can process per week, see our scaling guide.

Decision 2: Pacing — Waits Between Steps, Daily/Weekly Volume

Pacing determines how fast the sequence runs through prospects. Two layers of pacing decisions:

Within-prospect pacing. The wait days between each step for a single prospect. The shape above (3-5 days between connection and intro, 9-12 between intro and value message) is well-tested. Faster sequences — sending step 3 within 24 hours of acceptance, sending step 4 two days later — read as automation to LinkedIn's behavioral detection systems. Slower sequences (5-7 days between every step) work fine but extend the time-to-pipeline by 2-3 weeks per prospect.

Across-prospect pacing. The daily volume of new prospects entering the sequence per account. The locked operator number is 100 connection requests per week per floor-tier account, ~150/week for established accounts, ~200/week for ceiling-tier accounts. For the full tier breakdown, see our connection-limit explainer. Below that ceiling, the daily volume should be spread across business hours — sending 15 requests in one hour reads completely differently than sending 15 requests across 6 hours, even though the daily total is identical.

Cloud-based automation tools handle within-day pacing automatically when configured with realistic windows. HeyReach's default working-hours pacing distributes sends across the account's local timezone business day, which is usually correct out of the box.

Decision 3: Message Variants — Per-Account and Per-Segment Variation

Sequences fail at scale because they use identical templates across every account. Even when the template is well-written and personalized with custom variables (first name, company), if 30 accounts send the same template-skeleton message, LinkedIn's content systems pattern-match the structure and throttle every account simultaneously. For why this matters at the detection layer, see our restriction-detection guide.

Two layers of variation prevent this:

• Per-account variation. Each LinkedIn account in the sequence uses a different message template for the same step. If you have 10 accounts running the same outreach to financial-services prospects, each account uses a different 'Step 3 intro message' from a pool of 5-10 variants. The pool of variants is randomly assigned to each account, so no two accounts send identical sequences.
• Per-segment variation. Different prospect segments get different message angles. A 'CMO at SaaS company' prospect gets a different value-proposition message than a 'VP Sales at SaaS company' prospect, even from the same account. Most cloud-based tools support segment-based variants natively — you upload prospect lists with segment tags, and each segment has its own message-variant pool.

How HeyReach Handles This

HeyReach lets you assign multiple message variants per step and randomly distributes them across accounts and prospects. Five variants per step × four steps = effectively 625 unique sequence permutations across the campaign, which is enough variation that no two prospects receive identical messages even if dozens of accounts are running the campaign in parallel. Lemlist, Expandi, and La Growth Machine have similar variant systems with slightly different UIs.

Decision 4: Prospect-List Construction — Where Targets Come From

The prospect list is the campaign's input. A great sequence on a bad prospect list produces zero pipeline; a mediocre sequence on a great prospect list produces meaningful pipeline. Three approaches to building lists, in order of increasing quality:

• Sales Navigator search. Most outreach tools integrate directly with Sales Navigator searches — you build a search in Sales Nav, paste the URL into the automation tool, and the tool pulls the result set into the sequence. The advantage is freshness (Sales Nav results update as the platform's data updates) and depth of filtering (titles, seniority, company size, geography, industry, recent role changes).
• CSV import. You build the prospect list elsewhere (Apollo, ZoomInfo, a CRM export, a manually-built list) and upload as CSV. The advantage is control — you can validate the list before it enters the sequence, deduplicate against existing outreach, and apply business logic Sales Nav doesn't support.
• Trigger-based lists. Lists built from real-time signals — prospects who recently changed jobs, who recently posted about a relevant topic, who recently engaged with your content. These convert better than static lists because the message can reference the trigger naturally, but require additional tooling to build.

Quality Checks Before a List Enters the Sequence

Three checks every prospect list should pass:

Deduplication against existing outreach. Nothing kills trust faster than messaging a prospect who already received an identical sequence from another account in your operation last month. Most cloud-based tools have global deduplication across all your accounts — enable it.

Active-profile filter. Sales Navigator includes inactive accounts. Filter for prospects who logged in within the past 60 days; messages to dormant accounts waste sequence capacity.

Realistic match to the account. A US-based marketing manager sending outreach to senior executives at Fortune 500 manufacturing companies is implausible — the prospect won't accept, the acceptance rate drops, and the account's trust score follows. Match prospect seniority and industry to what your sender account plausibly represents.

Decision 5: Personalization Layer

Generic templated outreach gets accepted at 15-25%. Personalized outreach — with content that references something specific about the prospect or their company — gets accepted at 35-50%. The 15-25 percentage-point difference compounds across the whole sequence; campaigns that convert at 2x the connection rate convert at roughly 2x the reply rate at every subsequent step.

Three layers of personalization, in order of increasing effort:

• Custom variables. First name, company name, role title. Every automation tool supports these natively. Required minimum; adds maybe 5 percentage points to acceptance.
• Conditional content. Different message blocks for different prospect attributes — a CMO gets a different value proposition than a VP Sales, even within the same segment. HeyReach, Lemlist, and La Growth Machine support if-then conditional content in message templates.
• AI-generated per-prospect personalization. Tools that scan the prospect's recent posts, profile activity, or company news and generate a 1-2 sentence personalized intro per message. Adds significant per-message cost but produces the highest acceptance rates (often 50%+). HeyReach, Lemlist, and La Growth Machine have native versions of this; Clay-style external pipelines feed personalization data into many other tools.

Decision 6: Reply Handling

Replies are where the sequence ends and the human work begins. Two decisions about how this transition happens:

Where do replies land? Two options. The cloud-based tool can host a unified inbox — a view aggregating messages from all your LinkedIn accounts into one interface — where one team member handles replies across the entire account fleet. Or replies pipe into your CRM (HubSpot, Pipedrive, Salesforce, Close) via webhook, and your existing sales process handles them. Both work; the unified-inbox approach is operationally simpler for teams running campaigns end-to-end inside the automation tool, while CRM-piping fits teams whose sales process is already CRM-centered.

When does the sequence pause for a prospect? The default in every tool is 'pause on reply' — once a prospect responds, no further automated steps execute for them. Some tools also support 'pause on connection-request decline' or 'pause on opt-out keyword in reply.' Configure these correctly before launch; the worst case is a prospect who said 'not interested' getting follow-ups three days later because the sequence didn't stop.

Reply triage. Not every reply needs immediate human attention. Three categories: hot replies (interested, asking for more info, ready for a call) get same-day human response; warm replies (neutral, clarifying questions, soft objections) get a 24-48 hour response window; cold or negative replies get a polite acknowledgment and removal from sequence. A team running 25-50 accounts typically generates 30-70 replies per day — without triage, the operation drowns in low-value messages.

Decision 7: Tracking and Analytics

Most automation tools surface 15-20 metrics per campaign. Only four matter for operational decisions:

• Acceptance rate by account. If one account's acceptance rate is 20 percentage points below the others, something is wrong with that account specifically — sender profile, IP issues, or recent trust-score drop. Investigate before it cascades.
• Reply rate by sequence step. Tells you which message in the sequence is doing the work. Most campaigns get 60-70% of replies from one specific step; if it's step 1 (intro message), your value proposition is too weak in step 2; if it's step 4 (final follow-up), the earlier steps are over-pitching.
• Pipeline contribution rate. Of the replies you get, how many convert to actual sales conversations? Below 10%, your targeting is off. Above 30%, you're under-volume — the sequence is producing high-quality replies but at lower volume than the prospect list could support.
• Cost-per-reply. Total operational cost (accounts + tools + proxies + operator time) divided by qualified replies per month. The benchmark to beat: $20-40 per qualified B2B reply. Anything north of $80 means something in the stack is misconfigured.

Metrics That Don't Matter

Open rates (most LinkedIn messages don't have meaningful open-rate tracking), profile views generated (vanity metric), unsubscribe rates (rare on LinkedIn). Tools surface these because they're easy to measure, not because they correlate with pipeline.

HeyReach In Practice: The Full Sequence Setup

Putting the seven decisions together, here's how a complete sequence setup looks in HeyReach:

Setup step 1: Connect accounts. HeyReach connects to each rented LinkedIn account via the account's residential proxy and anti-detect browser session. Each account is a 'sender' in HeyReach's terminology. Rented accounts typically arrive with residential proxies bundled — the source we recommend is NextGen Profiles, which delivers real warmed-up accounts with dedicated residential proxies and recovery-first restriction handling included in the monthly fee.

Setup step 2: Build the sequence. In HeyReach's Campaign Builder, the seven decisions above map to seven setup screens: sequence shape (drag-and-drop step builder), pacing (wait-times between steps + daily-limit slider per account), message variants (multi-variant text editor per step), prospect list (Sales Nav URL paste or CSV upload), personalization (custom-variable picker + AI-Personalization toggle), reply handling (unified-inbox-on toggle + pause-on-reply default + opt-out keyword config), tracking (campaign-level dashboard surfaces the four metrics that matter).

Setup step 3: Stage the launch. New campaigns shouldn't launch at full volume on day one. HeyReach has a ramp-up setting that starts each account at 30-50% of target daily volume and increases over the first 2 weeks. This protects accounts during the initial 'is this campaign going to trigger restrictions?' window where any mistakes show up first.

Setup step 4: Monitor and adjust. After launch, the campaign dashboard surfaces account-level health (acceptance rates, reply rates, restriction events) and message-level performance (reply rates by sequence step). The first 2 weeks are when you tune variant performance, kill underperforming message templates, and adjust pacing if any accounts are showing detection signals. After that, the campaign typically runs steady-state with monthly check-ins.

How Other Tools Differ

Lemlist's multi-channel features layer email outreach on top of the LinkedIn flow, which fits teams running coordinated email + LinkedIn campaigns to the same prospects. Expandi is similar to HeyReach in single-channel focus but slightly more rigid in sequence-builder configuration. La Growth Machine's strength is the personalization layer — deeper integration with external data sources for per-prospect customization. Skylead and Dripify cover the same operational ground at lower price points; both are reasonable choices for solo operators or small teams running 1-5 accounts.

For a structured comparison of how the cluster recommends thinking about provider selection at different operational scales, see our roundup of LinkedIn account rental services.

Common First-Sequence Mistakes

Five mistakes that operators make on their first multi-account sequence:

• One template, all accounts. The biggest single failure mode. Even with custom variables filled in, identical template-skeletons across 20+ accounts get pattern-matched and throttled simultaneously. Always have 5-10 variants per step.
• Launching at full volume on day one. New campaigns should ramp from 30-50% of target volume over 2 weeks. Day-one full-volume launches produce 'mass-restriction events' where multiple accounts trigger detection together.
• Same sequence across all prospect segments. A 'CMO at enterprise SaaS' prospect and a 'Founder at 5-person startup' prospect get the same message, with the same value proposition, with the same close. Both ignore it. Segment your variants.
• No reply triage. Replies pile up unhandled because nobody's assigned to handle them. Hot replies go cold within 48 hours. The campaign produces pipeline only if someone responds to it in real time.
• Optimizing on the wrong metrics. Operators stare at acceptance rates and reply rates without checking cost-per-qualified-reply. Acceptance rates and reply rates are mid-funnel metrics; pipeline contribution is the only number that determines whether the sequence is worth running.