How to build an outreach sequence that converts in 2026

Last updated: June 2026 · Marcus Chen, outbound sales consultant, 12+ years in SDR leadership

---

TL;DR — 7 things to know before reading

• An outreach sequence is an architecture, not a series of emails: it has defined step count, defined timing between steps, defined message angles per step, and defined branch logic (what happens when someone replies, opts out, or does not reply)
• This guide uniquely covers sequence architecture: the structural decisions (how many emails, which angles in which order, when to send, what happens after each event) that determine whether a sequence produces 4% or 14% reply rates on the same ICP
• Step count matters less than step purpose: the most effective cold email sequences have 3–4 emails over 12–14 days, where each email serves a distinct purpose rather than repeating the same offer in slightly different words
• Per Instantly's cold email benchmarks, the average reply rate for cold email sequences is 8.5%; sequences with 3 substantive emails plus a break-up email (4 steps) outperform 2-step and 5+ step sequences at the same ICP
• The break-up email (Email 4, sent on Day 12–14) generates 15–25% of total positive replies in a 4-step sequence; most SDRs who run only 3-step sequences are missing this pipeline
• Branch logic is the most neglected element of sequence architecture: what happens when a prospect opens without replying 3 times (signal: engaged but not ready), or when a prospect unsubscribes (signal: wrong timing or wrong ICP), or when a prospect forwards to a colleague (signal: interested but not the decision-maker)
• Verdict: Instantly handles sequence execution, branch logic, stop-on-reply, and A/B testing; Quarvio provides the verified contacts the sequence runs against

---

Our take

Most cold email advice focuses on copy: what to write in Email 1, whether to use subject lines with questions or statements, how short emails should be. Copy is important, but it is downstream of architecture. A well-written Email 1 in a sequence with the wrong structure will underperform a mediocre Email 1 in a sequence with the right structure.

The unique angle of this guide is sequence architecture. Architecture covers the structural decisions that copy cannot compensate for: how many emails, how much time between them, what angle goes in which step, what happens after each recipient action (reply, no reply, open without reply, unsubscribe), and how to set up the sequence in Instantly to execute that architecture without manual intervention.

The guide covers the 4-step sequence architecture that produces above-average reply rates for most B2B ICPs, the branch logic that handles every reply scenario, and the A/B testing structure that improves performance over successive campaigns.

The 4-step sequence architecture

Why 4 steps (not 3, not 7)

Why not 3 steps: A 3-step sequence omits the break-up email (Email 4). Per benchmark data, the break-up email generates 15–25% of total positive replies in a well-structured sequence. Running only 3 steps leaves 15–25% of pipeline unmined.

Why not 5+ steps: Sequences longer than 4 substantive emails produce diminishing returns. After 4 emails over 14 days from an unknown sender, a prospect who has not replied has either: (a) seen the emails and decided not to engage, or (b) not seen the emails (deliverability issue). Adding steps 5, 6, and 7 primarily generates unsubscribes and spam complaints from group (a), which damages sending domain reputation.

Why 4 steps works: 4 emails allow 3 distinct message angles (Email 1, 2, 3) plus a pattern interrupt (Email 4 break-up). Each email serves a different purpose, which means each email is a genuinely different piece of communication rather than a variation of the same offer.

The 4-step structure

LinkedIn touchpoint: timed for Day 2–3 (after Email 1, before Email 2), run via Aimfox

Email 1: The problem email

Email 1 carries the highest leverage in the sequence. 40–50% of total positive replies come from Email 1 in a well-structured 4-step sequence. The structure:

Line 1 (the hook): Name the specific consequence of a specific problem the ICP experiences. This line should be so specific that a reader outside the ICP would not immediately recognise it, but someone inside the ICP would read it and think "that is exactly what is happening to us."

Wrong: "I help sales teams improve their pipeline."

Right: "Most [job title] I talk to at [company size] [industry] companies are [specific negative situation] because [specific root cause]."

Line 2 (the evidence): One proof point that validates the problem statement. Either a data point from an approved source or a specific result from a similar company.

Line 3 (the CTA): A soft, low-friction call to action. The correct CTA for Email 1 is not "book a demo" — the prospect does not know who you are yet. The correct CTA is: "Worth a 15-minute conversation to see if this is relevant to you?" or "Open to a quick chat about how we've approached this for similar teams?"

Email 1 formatting rules:

• Under 100 words (including the sign-off)
• Plain text only (no HTML formatting, no images, no headers)
• One CTA maximum
• No company description in the email body
• No attachment or link (in Email 1 only)

Email 2: The angle email

Email 2 is sent on Day 4 to all contacts who did not reply to Email 1. Email 2 serves two purposes: (1) it provides a second touchpoint for prospects who saw Email 1 but were not ready to reply, and (2) it presents a different frame on the same problem for prospects who were not moved by Email 1's angle.

Angle options for Email 2:

• Data contrast: "Companies in [industry] typically [negative benchmark]. The ones that [positive outcome] tend to [specific differentiating approach]."
• Problem reframe: Same problem as Email 1, but named from a different perspective (the department, the consequence for the company rather than the individual, or a longer time horizon).
• Objection pre-empt: Address the most common first objection. "I know most [title] already have [existing solution]. The difference we see is [specific gap that your approach fills]."

Email 3: The evidence email

Email 3 is sent on Day 8 to contacts who have not yet replied. Email 3's purpose is social proof: demonstrating that the approach works for companies similar to the prospect's company.

Evidence options for Email 3:

• Specific result from a named or implied company type: "One [similar company type] we worked with went from [before state] to [after state] in [timeframe]. The change was [what specifically changed]."
• Category-level G2 or review data: "We've got [X] reviews on [platform] from [company type] averaging [rating] for [specific benefit]."
• Benchmark comparison: "Industry average for [metric] in [category] is [number]. Our customers average [higher number]."

Note: Do not use the same type of evidence in Email 3 that you used in Email 1. If Email 1 used a data point, Email 3 should use a company result. Variety in evidence types provides more entry points for different prospect belief systems (some prospects are moved by data, others by peer results).

Email 4: The break-up email

Email 4 is the highest-ROI email in the sequence on a per-email basis. The break-up email generates 15–25% of total positive replies while requiring the shortest writing effort of any email in the sequence.

Break-up email structure:

Opening: "Didn't want to leave your file open without checking in."

Middle (choose one):

• "If this isn't relevant right now, no problem at all — just let me know and I'll close this out."
• "Should I take your silence as a no, or does it just mean timing is off?"
• "Last thing I want to do is clog your inbox if this isn't a fit. Worth a 5-minute call to find out, or should I close the loop?"

Close: "Either way, I appreciate your time."

Why break-ups work: The break-up email creates an implicit loss-aversion signal: the conversation is about to end. Some prospects who have been genuinely interested but too busy to reply feel motivated to respond rather than lose the conversation. The break-up also gives explicit permission to say no, which paradoxically increases yes rates by removing the pressure the prospect has been avoiding.

LinkedIn touchpoint architecture

Per Instantly's cold email benchmarks, multichannel outreach (email + LinkedIn) increases total reply rate by 40–60%. The LinkedIn touchpoint integrates into the 4-step email sequence at Day 2–3 via Aimfox.

LinkedIn connection note (under 200 characters): "Hi [First name], sent you an email about [topic] — dropping a connection here in case this is easier. [Your name]"

LinkedIn follow-up (after connection accepted): Brief, no-pitch: "Thanks for connecting, [First name]. Happy to share more about [one specific thing] if useful — no pressure."

LinkedIn branch logic:

• If LinkedIn connection is accepted + they reply on LinkedIn: suppress from email sequence immediately, handle in LinkedIn
• If LinkedIn connection is accepted + they don't reply on LinkedIn: continue email sequence, no LinkedIn follow-up needed
• If LinkedIn connection is not accepted by Day 10: no further LinkedIn action

Branch logic architecture

Branch logic is the set of rules that determine what happens to a contact based on their actions during the sequence. Poorly configured branch logic is the second-most-common cause of sequence underperformance (after Email 1 angle quality).

Branch logic table

The stop-on-reply configuration

Stop-on-reply is the most important sequence configuration setting in Instantly. When enabled, any reply from a contact (positive, negative, or out-of-office) pauses the sequence for that contact. Without stop-on-reply, a contact who replies "Not interested" will continue to receive Emails 2, 3, and 4, generating additional spam complaints and damaging domain reputation.

Verify in Instantly: Campaign settings → Stop sending on reply: ON. Never disable this setting.

Out-of-office handling

A significant portion of "replies" in any campaign are out-of-office autoresponders. These replies trigger stop-on-reply (the sequence pauses), but the contact has not actually responded. Options:

• Option A: Rely on the contact to re-engage when they return (passive)
• Option B: After 3 days, manually re-activate the sequence for confirmed out-of-office contacts (active)

Instantly provides an "out of office" detection label that distinguishes autoresponder replies from genuine replies. Review the out-of-office label daily and decide per contact whether to re-enable the sequence.

Sequence configuration reference

Sequence A/B testing framework

What to A/B test (and in what order)

Test 1: Email 1 opening angle (highest leverage)

Run 2–3 variants of Email 1's first line, each naming the problem from a different angle or consequence. Keep everything else identical (Email 2, 3, 4, subject line). Measure: open rate (subject line signal) and reply rate (Email 1 angle signal) after 200+ sends per variant.

Test 2: Subject line (after Email 1 angle is validated)

Run 2 variants of the subject line for the winning Email 1 angle. Common subject line patterns to test:

• Pattern A: First name only ("Quick question, [firstName]")
• Pattern B: Role-specific reference ("For [title] teams like yours")
• Pattern C: The specific problem in 5–7 words

Test 3: Break-up email framing

After Emails 1–3 are validated, test 2 variants of the break-up email. Some ICPs respond better to the explicit "should I close your file?" framing; others respond better to the softer "worth 5 minutes?" framing.

A/B testing rules

• Test one variable at a time only. Never test Email 1 angle and subject line simultaneously — the improvement cannot be attributed to either change.
• Minimum 200 sends per variant before declaring a winner. Below 200 sends, reply rate variance is noise, not signal.
• The winning variant becomes the default; the losing variant is retired. Do not run indefinitely — the test goal is to identify and commit to a winner.

Advanced sequence tactics

Tactic 1: Write Email 4 first, then work backwards

The break-up email is the clearest expression of what the sequence is trying to accomplish. Write the break-up first ("we're trying to get a yes or a definitive no by Day 14") and then work backwards to write Emails 1, 2, and 3 as the journey toward that break-up. This approach produces a more coherent narrative arc across the sequence than writing Email 1 first and trying to build toward a break-up that was not planned.

Tactic 2: Use the Email 3 evidence email as an ICP match test

The specific company type or result referenced in Email 3 is a filter for ICP quality. If Email 3 references "SaaS companies in our segment" and 20% of replies to Email 3 say "we're not SaaS" or "that's not us," the contact list includes significant non-ICP contacts. This is a useful diagnostic: Email 3 double-checks ICP quality by referencing a specific relevant peer group. If ICP match rate in Email 3 replies is below 80%, tighten the Quarvio filter criteria for the next contact list order.

Tactic 3: Create an "intent signal" variant for Email 2

For contacts who opened Email 1 (as tracked in Instantly open rate) but did not reply, write a different Email 2 than for contacts who did not open Email 1. The "high-intent" variant (for openers) references the Email 1 message: "Following up on my note about [topic] from [X] days ago — wanted to make sure you had context on [specific piece of evidence]." This creates continuity for engaged-but-not-yet-ready prospects. Instantly's segmentation can filter contacts by open status for conditional Email 2 routing.

Tactic 4: Time Email 2 to land on a different day of the week from Email 1

If Email 1 lands on a Monday, time Email 2 to land on Thursday or Friday. The reason: a prospect who didn't reply to Email 1 on Monday may have been busy, in meetings, or dealing with a specific Monday priority. Email 2 on Thursday reaches them at a different point in their week, increasing the chance of engagement. Varying delivery days within the sequence is a simple configuration change in Instantly's send schedule that produces a modest but consistent improvement in Email 2 reply rate.

Tactic 5: Build a sequence library indexed by ICP and problem type

After running 5+ campaigns, the agency or SDR team accumulates validated sequence patterns: specific Email 1 angles that work for specific ICPs, specific Email 3 evidence formulations that resonate with specific industries, and specific break-up framings that work for specific seniority levels. Index these validated patterns in a shared document organised by ICP type, problem type, and email step. New campaign sequences can then be assembled from proven building blocks rather than written from scratch, reducing copy writing time and improving first-campaign performance.

Troubleshooting sequence architecture

Problem 1: High open rate but low reply rate

Symptom: Open rate is 35%+ (good) but reply rate is below 4% (poor).

Cause: The subject line is working (prospects are opening the email) but Email 1 is not resonating. Either the problem statement is too generic (the ICP recognises the category but not the specific consequence), or the CTA asks for too much commitment too early ("30-minute demo" in Email 1 from a cold sender reduces reply rate significantly).

Fix: Rewrite Email 1. Specifically: make the opening problem statement more specific (name a daily consequence, not a category-level problem) and replace the CTA with a lower-friction question ("Is this something worth exploring?" not "Can we schedule a demo?").

Problem 2: Email 1 performs well but Emails 2 and 3 add no replies

Symptom: 60%+ of positive replies come from Email 1 only. Emails 2 and 3 generate almost no additional responses.

Cause: Emails 2 and 3 are restating the same angle as Email 1 in different words. They are adding no new information or perspective for the prospect.

Fix: Review Emails 2 and 3. If they are variations of the same problem statement from Email 1, rewrite them with genuinely different angles or evidence types. Email 2 should provide information that a prospect who was interested in Email 1 would find compelling to read. If they are already in different enough, the ICP may not have a "second problem" that makes additional follow-up relevant — in that case, a 3-step sequence may be more appropriate than 4.

Problem 3: Break-up email (Email 4) generating most replies but the replies are negative

Symptom: 60%+ of all replies come from Email 4, but most are "Please remove me" or "Not relevant."

Cause: The break-up email is effective at prompting responses, but the ICP or offer is not resonating, so the responses are opting out rather than engaging.

Fix: If Email 4 is driving "please remove" responses at high rates, the sequence is reaching the wrong people (ICP issue) or the offer is not positioned for the specific problem the ICP experiences (copy issue). Do not change the break-up email — fix the upstream issues. Review the Quarvio filter criteria for ICP tightness, and review Email 1 for offer specificity.

Problem 4: Stop-on-reply is enabled but sequence continued to a replied contact

Symptom: A contact replied and the sequence continued, sending them Email 2, 3, or 4 after their reply.

Cause: Stop-on-reply was enabled but the reply was an out-of-office autoresponder that the system re-activated after the OOO expiry, or the reply came from a secondary email address that is not matched to the contact record in Instantly.

Fix: Check the Instantly activity log for the contact. If the sequence continued after a genuine human reply (not OOO), contact Instantly support. If the sequence resumed after an OOO expiry, check the "resume sequence after out-of-office" setting in Instantly and adjust the OOO resume behavior.

Problem 5: A/B test running for 6 weeks but still no clear winner

Symptom: Two Email 1 variants running for 6 weeks and 800+ sends per variant, but reply rates are within 0.5% of each other.

Cause: The two variants are not different enough in their approach for the reply rate difference to be meaningful. They may be testing two similar framings of the same problem.

Fix: Declare the test inconclusive and pick one variant to continue with (the one with slightly higher reply rate, or the one the team prefers for future iteration). Write new variants that test genuinely different Email 1 structures: one that names the problem first versus one that opens with social proof, or one that uses a data point versus one that uses a consequence-first narrative. The next test should have maximum contrast between variants.

Problem 6: Sequence is performing well on email but LinkedIn touchpoints are reducing email reply rate

Symptom: After adding Aimfox LinkedIn touchpoints, total positive reply rate dropped from 9% to 7%.

Cause: The LinkedIn connection requests are arriving before Email 1 (wrong timing), or the LinkedIn message is duplicating the Email 1 problem statement and creating a "double-pitch" impression that makes prospects less likely to reply to email.

Fix: Check Aimfox campaign timing: the connection request should arrive on Day 2–3 (after Email 1, before Email 2). If timing is correct, review the LinkedIn connection note: it should not pitch the product. The note should only identify who you are and why you're connecting. The email sequence carries the offer; LinkedIn carries the relationship signal.

Problem 7: Sequence is working for one ICP but the same architecture is failing for a new ICP

Symptom: 4-step sequence producing 12% reply rate for ICP A. Same sequence architecture (with new copy) producing 4% reply rate for ICP B.

Cause: Different ICPs require different sequence architectures, not just different copy. ICP B may require more evidence in an earlier step (put social proof in Email 2 instead of Email 3), or may require a shorter sequence (3 steps instead of 4, because the buying cycle for the ICP is shorter), or may respond better to a data-led Email 1 instead of a consequence-led Email 1.

Fix: Treat ICP B as a new architecture test, not a copy revision. Run the standard 4-step architecture for 300+ sends to establish a baseline, then identify specifically which step is underperforming. Make one architectural change (move the social proof, change the CTA format, adjust the step timing) and run for another 300+ sends. Document the architectural differences that emerge for this ICP for future campaigns.

Problem 8: Personalisation variables showing raw placeholder text in delivered emails

Symptom: Prospects reply saying they received an email that said "Hi , your company " (empty variables).

Cause: The contact list had empty fields for firstName or companyName, and Instantly delivered the email with the empty variable placeholder.

Fix: Before any campaign launch, run a "variable completeness check" on the contact import: verify that all personalisation fields used in the sequence are populated for 100% of contacts. Remove any contacts with missing personalisation fields before import. Add a fallback value to the sequence for critical variables: "there" as a fallback for firstName if the name is missing.

Community evidence

Instantly's cold email benchmarks analysed thousands of B2B cold email sequences and found that 3–5 step sequences consistently outperform shorter (1–2 step) and longer (6+) sequences. The study attributes this to the "follow-up reply premium": contacts who engage with a sequence (open without replying) are increasingly likely to reply as they see the follow-up emails — up to a point, after which additional touches produce diminishing returns and increasing negative responses. The 4-step architecture sits at the optimal point on this curve.

Instantly's 2026 benchmark report reports that the top 10% of sequences (above 10% reply rate) share a common structure: a short, specific Email 1 (under 100 words), a follow-up sequence with genuinely different angles per step, and a break-up email at Day 12–14. The report confirms that length and step count are secondary to angle distinctiveness: a 4-step sequence with 4 variations of the same angle underperforms a 3-step sequence with 3 genuinely distinct angles.

"The single best thing we did to our sequences was stop treating Email 2 and 3 as reminders of Email 1 and start treating them as independent arguments. Email 1 makes one argument. Email 2 makes a completely different one. Email 3 makes a third. The prospect who wasn't moved by argument 1 might be moved by argument 3. Our reply rate went from 7% to 13% with the same contact list after that change." — G2 reviewer, Instantly reviews on G2

Our actual stack

Frequently asked questions

How long should each email in the sequence be?

Email 1: under 100 words. Emails 2 and 3: under 80 words each. Email 4 (break-up): under 60 words. Longer emails are not read more carefully — they are read less. At under 100 words, a mobile viewer can read the entire email without scrolling. Above that, engagement drops with each additional sentence. The 4-step sequence should have a total word count of under 320 words across all four emails.

Should I include links in every email?

No. Email 1 should have no links (links increase the probability of spam filter flagging for a sender the recipient does not yet know). Email 2 and 3 can include one link each (case study, resource, or company website) as evidence, but not as the primary CTA. Email 4 should not include links. The CTA across all emails is a reply or a calendar booking — not a link click.

What is the best day to send Email 1?

Tuesday and Wednesday show the highest reply rates in most B2B research. Monday is high-volume inbox day; Thursday and Friday show declining engagement as the week progresses. Monday and Friday are the lowest performers. Set Instantly to send Email 1 on Tuesday or Wednesday in the prospect's timezone.

How do I handle it when a prospect asks for more information via email?

A prospect who replies with "can you send me more information?" is expressing interest but not committing. The correct response is not to send a 3-page brochure — it is to interpret the request as a soft signal of interest and propose a brief call: "Happy to walk you through it — easier to share what's most relevant in 15 minutes. Does [time option 1] or [time option 2] work?" This converts the information request into a meeting more effectively than email-based information sharing.

Should I personalise every email in the sequence or just Email 1?

Personalise Email 1 most heavily (custom first line, ICP-specific problem statement). Emails 2 and 3 should include the prospect's first name and company name in the opening line, but the body can be largely templated. Email 4 (break-up) is typically the least personalised email in the sequence. The ROI on personalisation is highest in Email 1 because that is where the reply decision is most often made.

How do I set up A/B testing in Instantly?

In Instantly, create a campaign and enable the A/B test toggle. Upload the contact list once; Instantly distributes contacts between variants automatically. Set the variant split (50/50 for 2 variants, 33/33/33 for 3 variants). Monitor results in the campaign analytics dashboard. After 200+ sends per variant, Instantly shows reply rate, open rate, and bounce rate by variant. Manually select the winner and continue the campaign with only the winning variant.

What happens to the sequence when a contact books a meeting directly via a calendar link?

If the CTA in any email includes a calendar link and the prospect books directly (without replying to the email), the sequence may continue (Instantly does not know a meeting was booked). Configure a Zapier integration: when a meeting is booked in the calendar tool, trigger an Instantly contact suppression or sequence pause for the contact. Without this integration, prospects who book via calendar link may continue to receive follow-up emails even after booking.

Can I run the same sequence to the same contact list twice if the first run didn't produce enough results?

No. Running the same sequence to the same contact list twice (or more) dramatically increases spam complaint rates and unsubscribes. Contacts who received 4 emails from you and did not reply are signalling disinterest or irrelevance. Re-contacting them with the same offer damages domain reputation. If the contact list is relevant, refresh it with new contacts from Quarvio. If the ICP is valid, write a new sequence with a different Email 1 angle and run it to a new contact list.

---

Sequences are only as effective as the contacts they reach.

Quarvio delivers pre-verified B2B contacts filterable by job title, company size, and industry — so the sequence runs against an ICP-matched, deliverable contact list. One-time purchase, credits valid for 12 months.

Browse Quarvio contact packages →