Contents
- Fast answer
- Quick facts
- What this covers
- Why it matters
- Where to go next
- Outcomes
- Step by step
- The methodology: how the sequence is assembled
- The three things you must add before the sequence runs
- A worked example: an HVAC company chasing system quotes
- How many touches it really takes in 2026
- The multichannel evidence: why sequences mix channels
- What changed in 2025-2026: volume stopped working
- The limits: a schedule is not a sequence
- When not to use this tool, and what to use instead
- At a glance
- Mistakes to avoid
- FAQ
- Related guides
Fast answer
The sales sequence builder creates cold outreach, warm follow-up, and re-engagement sequences with day-by-day templates for email, LinkedIn, and phone, plus AI automation suggestions. Every sequence follows a common structure: a pattern interrupt on the first touch, value delivery on touches two and three, social proof on touch four, and an urgency close through touches five to seven, with messaging customized to your sequence type and channel mix. It maps each touchpoint to the channel where it performs best at that position, using LinkedIn for personal initial interrupts, email for detailed value propositions, and phone calls for urgency at the close, Cadence follows engagement psychology: early touches cluster to build recognition before the prospect forgets you, later touches widen to respect their time, and re-engagement sequences use wider gaps to avoid appearing desperate. Personalization layers trigger events, industry challenges, role-based pain points, and tech stack, while AI handles research, enrichment, scheduling, and tracking rather than the relationship itself.
Facts last verified 2026-07
Quick facts
- Avg touches to get a response
- 4.81 (Belkins)
- Avg touchpoints to book a meeting
- About 8; top performers 5
- Deals requiring 5+ touches
- 80%
- Reps who quit after one touch
- 44%
- Avg cold email reply rate
- About 3.8%
- Multichannel (3+ channels) lift
- Up to 287% more responses
On this page
What this page covers
A tool visitor should leave with a decision, not just a number: build now, prepare first, choose another workflow, or follow a course path.
- 01Sequence type
- 02Channels
- 03Templates
- 04Automation
Why does this matter now?
Every sequence follows a common structure: pattern interrupt (touch 1), value delivery (touches 2-3), social proof (touch 4), and urgency close (touches 5-7). The specific messaging within each slot is customized to your sequence type and channel mix. The templates are our own starting drafts, not measured winners, so track your own reply rates and edit.
Internal path
Where to go next from this page
These links are part of the A8gent learning and conversion path. Use them to move from concept, to diagnosis, to workflow build, to course.
What you should be able to do after this
- Multi-touch sequences
- Channel orchestration
- AI automation suggestions
- Template library
How do you do it, step by step?
1. Sequence Architecture
Every sequence follows a common structure: pattern interrupt (touch 1), value delivery (touches 2-3), social proof (touch 4), and urgency close (touches 5-7). The specific messaging within each slot is customized to your sequence type and channel mix. The templates are our own starting drafts, not measured winners, so track your own reply rates and edit.
2. Channel Orchestration
Multi-channel sequences outperform single-channel by 25-40%. Our builder maps each touchpoint to the channel where it performs best at that position in the sequence. LinkedIn works for initial pattern interrupts because it is personal. Email carries detailed value propositions because it allows length. Phone calls create urgency at the close because they demand immediate response. The channel order is optimized by audience type.
3. Cadence Optimization
Spacing between touches follows engagement psychology research. Early touches are clustered (days 1, 3, 5) to build name recognition before the prospect forgets you. Later touches widen (days 8, 12) to respect their time without losing momentum. We adjust cadence by sequence type because warm follow-ups tolerate faster pacing while re-engagement sequences need wider gaps to avoid appearing desperate.
The methodology: how the sequence is assembled
Three inputs build the schedule. You pick a sequence type from cold outreach, warm follow-up, re-engagement and event follow-up. You set the number of touches between 3 and 8. You select any combination of email, LinkedIn and phone. The tool then pairs a day number with a channel and a stored message for each touch.
The day spacing is a fixed array, not a calculation. Cold outreach uses days 1, 3, 5, 8, 12, 16, 21 and 28. The other three types use days 1, 2, 4, 7, 10, 14, 18 and 23. Your touch count simply takes the first N entries, so a five-touch cold sequence runs on days 1, 3, 5, 8 and 12 and finishes inside a fortnight. The widening gaps are the design: early touches are close enough to read as one conversation, later ones far enough apart to read as persistence rather than pestering.
Channels round-robin in the order you selected them. Touch one uses your first channel, touch two your second, and so on, cycling. Two channels alternate; three channels rotate. There is no logic about which channel suits which moment, so if you want a phone call at touch three rather than touch two, you reorder your selection or edit the output.
The message library holds three stored messages per sequence type per channel, and the tool cycles through them by touch index. That has a consequence worth knowing: on a sequence with more than three touches on the same channel, the fourth message on that channel repeats the first one verbatim. Touch four is not a new message, it is message one again. Treat the library as three worked examples of an opener, a middle touch and a breakup, and write the rest yourself.
The three things you must add before the sequence runs
The stop condition first, because it is the difference between a sequence and a nuisance. Any reply pauses the sequence immediately, and so does the outcome you were chasing: a booking, a payment, an accepted quote, a document received. The stop should be detected from the system of record rather than from somebody remembering to switch it off. Nothing in the builder expresses this, and it is the single most important line in the specification.
The suppression list second. One authoritative list of people who have asked not to be contacted, honoured across every channel and every tool you send from. A person who opted out of email and then receives a text has not received a clever multi-channel touch, they have received evidence that you do not keep records. Opt-out handling has to be the same fact in every system, which usually means one system owns it and the others read it.
The consent and channel rules third. Email, phone and SMS have different requirements and different jurisdictions, and adding a channel to a sequence adds its rules along with it. In the United States, moving these touches to SMS means A2P 10DLC registration: Twilio's documentation sets out a brand step identifying who is sending and a campaign step describing the message purpose and how recipients opt in, opt out and get help, states that unregistered 10DLC traffic attracts additional carrier fees, and caps daily throughput by brand type, with sole proprietor brands limited to 1,000 SMS segments and MMS per day to T-Mobile.
Then add the two things that make any of it work: a trigger that fires the sequence from a real event in a real system, and a personalisation source. A sequence that starts because somebody imported a list behaves differently from one that starts because a quote reached day two without a status change, and the second is worth many times the first.
A worked example: an HVAC company chasing system quotes
An HVAC and plumbing company picks warm follow-up, five touches, email and phone. The builder returns touches on days 1, 2, 4, 7 and 10, alternating email, phone, email, phone, email, with three stored messages per channel cycling so the fourth phone script would repeat the first. Reasonable spacing for a warm lead, and too fast for a system replacement quote.
The real cadence for high-value trades quotes runs over two to three weeks rather than ten days, because the customer is getting two other quotes and talking to a partner about money. Our page for HVAC and plumbing describes the shape, and the unsold estimate follow-up guide gives the working version: an estimate-state trigger, five touches of real copy, stop-on-reply and stop-on-acceptance detected from the field service system itself rather than from a person.
The content of the middle touches is where the trades version diverges most. A generic middle touch offers a case study; a good one answers the three questions that actually stall a system replacement decision - financing, warranty and rebate eligibility - because those are the reasons the customer went quiet. The builder cannot know that. The industry page and the guide can.
Finally, the high-value branch. Above a certain quote value the right fifth touch is not a message at all, it is a task raised for a human to phone. Automating the small quotes end to end and escalating the large ones to a person is the design that consistently works in the trades, and no fixed spacing array will produce it for you.
How many touches it really takes in 2026
The persistence data justifies every follow-up slot in your generated sequence. Belkins' follow-up research finds an average of 4.81 touches required to get a response regardless of lead temperature, and broader outbound benchmarks show booking a first meeting takes about 8 touchpoints on average, with top performers getting there in 5. Meanwhile 80 percent of deals require five or more touches, yet 44 percent of reps give up after a single attempt, which is the single most fixable gap in most outbound motions.
Reply expectations should be set against current averages, which fell as inboxes filled with AI-written outreach. Recent benchmark aggregations put average cold email reply rates near 3.8 percent with meeting booking rates around 0.8 percent. The spread around the average is the real story: analyses find senders using relevant, insight-led personalization reaching up to 18 percent reply rates while generic template blasts sit near 5 percent or below.
Read your sequence output with those numbers in mind. A seven-touch cold sequence hitting a 4 to 6 percent reply rate with reasonable targeting is performing at or above benchmark; expecting 20 percent from a cold list is fantasy, and consistently below 1 percent signals a list or deliverability problem, not a copy problem. Fix problems in that order: list quality, deliverability, offer, then copy, because the first three cap what the fourth can achieve.
The multichannel evidence: why sequences mix channels
Channel mixing is the highest-leverage structural choice in sequence design, and the effect sizes are large. Outreach benchmark studies find multichannel sequences using three or more channels deliver up to 287 percent more responses than single-channel outreach, and multi-channel cadences combining email, LinkedIn, and phone outperform single-channel sequences by two to three times on reply rate. The mechanism is simple: each channel catches people the others miss, and repeated familiar names get warmer responses on every channel.
LinkedIn-specific benchmarks explain its position early in generated sequences. Current outreach data shows connection request acceptance rates around 27 percent and post-connection reply rates around 11 percent, with soft-touch LinkedIn nurturing reaching reply rates near 11.9 percent, higher than any pure email sequence in the same dataset. LinkedIn touches work best as familiarity builders that lift subsequent email response, which is why sequences front-load them rather than saving them for the close.
Phone remains the urgency channel and earns its late-sequence slot. By the fifth or sixth touch, a prospect who has seen your name on two channels is dramatically more likely to take a call than a pure cold call target. For teams that cannot staff calling, a voice note on LinkedIn or a short Loom video fills the same slot: a higher-effort, higher-signal touch that breaks the pattern of text in an inbox.
What changed in 2025-2026: volume stopped working
Two forces compressed outbound economics at once. Gmail and Yahoo bulk sender rules made high-volume, low-quality email operationally risky, with authentication requirements and spam complaint thresholds that throttle spray-and-pray senders. Simultaneously, AI writing tools made passable personalization free, which means passable personalization no longer differentiates. The combined effect shows up in the benchmark drift: average reply rates fell while the gap between average and well-targeted senders widened.
The response that works is research-heavy, volume-modest sequencing. AI agents now handle the research layer economically: tools like Clay and Apollo enrich prospects with trigger events, tech stack, and role signals, and workflows in n8n or Make.com can route that data into sequence tools automatically. The winning division of labor keeps humans on offer design and reply handling while agents do enrichment, drafting for review, send-time optimization, and engagement-based branching, such as accelerating the cadence when a prospect clicks twice.
Decision guidance for your generated sequence: pick the smallest list you can tolerate and the most research you can afford, because the benchmarks say a 200-prospect list at a 10 percent reply rate beats a 2,000-prospect list at 1 percent, and only one of those approaches damages your domain. Then instrument the sequence: track replies by touch and channel, and prune touches that never produce replies rather than adding more. Sequences improve by deletion more often than by extension.
The limits: a schedule is not a sequence
The builder produces days and channels. It has no trigger, no stop condition, no branching, no suppression list, no consent record and no CRM. Everything that makes a sequence safe to run in a real business is outside the tool, which is fine as long as you know it. The output is a plan for a conversation, not a system.
The message library is three examples deep per type and channel, and it repeats from touch four onward on any channel used more than three times. It also contains bracketed placeholders throughout, and those placeholders are the message: a template sent with [pain point] intact is not a mistake in copy, it is evidence of not reading before sending.
There are no performance figures anywhere on this page, and that is deliberate. We publish no reply rates, connect rates or optimal touch counts for these sequences, because we have not measured them and outbound performance depends far more on list quality, timing and offer than on the number of touches. Any figure you adopt should come from your own sends, on your own segment, counted over a defined period.
And the tool takes no view on whether you should be sending at all. Cold outreach has different legal and practical constraints from following up an existing customer, and the same schedule is offered for both. Deciding which of your lists you may contact, on which channels, is a question about your own records and your own jurisdiction, and it belongs before the spacing question rather than after it.
When not to use this tool, and what to use instead
Do not use it to plan a single message. If you need one email, the email template library gives you fifteen shapes with the personalisation prompts attached, and it is a faster route to a sendable draft.
Do not use it for the reminder and notification family. Appointment reminders, invoice chasing and review requests all look like sequences and behave differently: their triggers come from a system event, their stop conditions are non-negotiable, and their timing is dictated by the appointment, the due date or the job. The reminder guide, the QuickBooks invoice chasing ladder and the automated review request guide cover those properly.
Do not use it as the specification you hand to a builder. The trigger, source of truth, review rule and output destination are what a builder needs, and none of them appear here. The workflow planner produces that artefact in about half an hour, and the sequence then slots into it as the message layer.
And if the sequence will run on SMS to US recipients, the registration work comes first. Read the A2P 10DLC registration guide before you write a word of copy, because the brand and campaign path you choose sets your daily ceiling and the timeline is measured in days rather than minutes. If you would rather have the whole thing built, registered and handed over with the stop conditions wired to your system of record, that is what the done-for-you route covers.
At a glance
Outbound sequence benchmarks from 2025-2026 studies
| Metric | Benchmark | Source |
|---|---|---|
| Cold email reply rate | About 3.8% average | 2025 outbound benchmark aggregations |
| Insight-led personalized outreach | Up to 18% reply | Sales benchmark analyses |
| Meeting booking rate | About 0.8% of sends | 2025 outbound data |
| LinkedIn connection acceptance | About 27% | Cold outreach benchmarks |
| LinkedIn reply after connection | About 11% | Cold outreach benchmarks |
| Multichannel vs single channel | 2 to 3x reply rate | Multichannel studies |
What mistakes should you avoid?
- Lead with a trigger event: Reference something that just happened to the prospect: a funding round, a job change, a product launch, a conference appearance.
- Keep touch one under 80 words: Your first message competes with dozens of others in their inbox.
- Vary your ask across touches: Do not repeat the same CTA in every message.
- Use AI for research, not relationships: Let AI agents handle prospect research, data enrichment, send scheduling, and engagement tracking.
FAQ
- How does the sales sequence builder decide the timing?
- From a fixed array rather than a calculation. Cold outreach uses days 1, 3, 5, 8, 12, 16, 21 and 28. Warm follow-up, re-engagement and event follow-up all use days 1, 2, 4, 7, 10, 14, 18 and 23. Your touch count takes the first N entries from that array, so a five-touch cold sequence ends on day 12. The widening gaps are the point: early touches read as one conversation, later ones read as persistence rather than pestering. Adjust the tail if your buying cycle is longer than the array assumes.
- How are channels assigned to each touch?
- By round-robin, in the order you selected them. Touch one uses your first channel, touch two your second, and the rotation repeats. There is no logic about which channel suits which moment, so if you want the phone call at touch three rather than touch two, change the selection order or edit the output. In practice a good mix front-loads the low-friction channel and saves the call for the moment when there is something specific to say, which is a judgement the tool does not make.
- Why does touch four repeat the wording of touch one?
- Because the library holds three stored messages per sequence type per channel and cycles through them by touch index. On any channel used more than three times, the fourth message repeats the first verbatim. Read the three as worked examples of an opener, a middle touch and a breakup, and write the rest yourself. That is not a bad outcome: the later touches are where your own specifics do the most work, and a recycled opener at touch four is exactly the message a prospect notices for the wrong reason.
- What has to be added before a sequence can run safely?
- Three things. A stop condition detected from your system of record, so any reply or the outcome itself - a booking, a payment, an accepted quote - halts the sequence without anyone remembering to. One authoritative suppression list honoured across every channel and tool, because an opt-out respected in one system and ignored in another is worse than no automation. And the consent rules for each channel you have added. None of those exist in the builder, and all three are the difference between a sequence and a complaint.
- What do I need to do before sending a sequence by SMS in the US?
- Register for A2P 10DLC first. Twilio's documentation describes it as the standard US carriers put in place so that traffic to US end users through long code numbers is verified and consensual, and it has two steps: a brand identifying who is sending, and a campaign describing the message purpose and how recipients opt in, opt out and get help. Twilio states that sending from a 10DLC number without registering attracts additional carrier fees, and daily throughput is tiered by brand type, with sole proprietor brands capped at 1,000 SMS segments and MMS per day to T-Mobile.
- How many touches should a sequence have?
- We do not publish an optimal number, because we have not measured one and the honest answer depends on your buying cycle rather than on a benchmark. What is reliable is the shape: enough touches that a single missed message does not end the conversation, spacing that widens over time, a genuinely different reason to make contact each time, and an explicit final message that closes the loop gracefully. If you cannot think of a distinct reason for touch six, that is your answer for how many touches you should have.
- Should AI write the messages in the sequence?
- Use it for the research and the drafting, and keep a person on the send until you have evidence. The reliable division is that automation handles the trigger, the timing, the stop condition, the logging and the first draft, while a human owns the claim being made and the decision to send. Draft mode for the first month, with somebody counting how often they edit before sending, gives you an error rate you can quote. Autonomy on outbound messaging is something a workflow earns from that number, not something it starts with.
Sources & further reading
Was this page helpful?
