Sequence architecture
TL;DR
- Default: 3-5 emails counting the opener, widening gaps — day 1, 3, 7, 14 — rotating the angle each touch.
- Follow-ups do most of the work: 55-65% of all replies arrive after the first email across the large data reports (~2024, ongoing).
- Since the 2024 bulk-sender rules, every added touch is spam-complaint surface; the long-cadence playbooks predate that reality.
Touch count is cold email's loudest disagreement, and each camp is honest about a different constraint: attention, deliverability, or effort.
The schools
High-touch (Sam Nelson's Agoge Sequence, ~2018-19 [1, 2, 3]) — Around 15 touches over roughly a month, across email, phone, and social. Day 1 is a triple touch — hyper-personalized email, call, and connection request — and later emails are short threaded bumps riding the day-1 email's quality; a breakup closes it (Breakup emails). Born on an SDR team working small lists of high-value accounts, it assumes that staffing.
Fits when: high-value accounts, small lists, a team that can genuinely work phones and social. · Avoid when: solo senders, large cold lists, no calling motion — thinned out, it degrades into pestering.
Recognizable by: in the email trail, "Re:" bumps stacked on one strong opener and a breakup at the end; beyond email (invisible in the trail), a day-1 multichannel burst and calls between the emails.
Few-touches (the complaint-economics camp, 2024-ongoing) — The argument is arithmetic more than corpus: once the 2024 bulk-sender rules made spam-complaint rate an existential metric, every touch past the point of diminishing returns buys little reply and real complaint surface. Be honest about the evidence: no current large corpus backs a hard two-email cap — the 2026 follow-up study of 7.5M emails [4] names 3-5 steps the practical sweet spot, while still measuring the tail decay this camp argues from (per-step value falls sharply past step 5; step 6 replies at 0.28%, 2026). Read it as cap discipline: stop where the marginal reply no longer pays for the marginal complaint risk — on complaint-sensitive lists that can mean two or three touches.
Fits when: deliverability is the binding constraint — large lists, complaint-sensitive niches. · Avoid when: small lists of high-value accounts where measured persistence is expected.
Recognizable by: sequences ending after email two or three, no breakup note, most of the effort concentrated in the opener.
Moderate 3-5 (the data-report mainstream, 2023-2026) — The corpora converge on the shape even where they disagree on the share. A 20M+ email industry study (2026) [5] reports sequences with 3-5 follow-ups replying at 8.3% against 4.1% with none, the first follow-up contributing most at around 8.4%; a 65M-journey report (2023) [6] puts 55% of all replies on follow-ups; and a 7.5M-email 2026 study names 3-5 steps the practical sweet spot, with 58.6% of replies arriving on follow-ups [4]. The one caveat worth carrying: the 2026 industry study also attributes 58% of replies to the first email, contradicting the follow-up shares above — the share is disputed (Benchmarks), the doubling is not. The workhorse shape: 3-5 emails counting the opener, widening gaps (day 1, 3, 7, 14), a different angle each touch — the same shape Corey Haines' [7] and LeadMagic's [8] open-source skill packs (2025-26) teach.
Fits when: almost everything — the sane default across list sizes and offers. · Avoid when: complaints are already elevated; trim toward few-touches.
Recognizable by: 3-5 emails, gaps that widen, each email standing on a fresh angle, a polite closer.
One caveat outranks all three schools: long cadences were designed before the 2024 bulk-sender rules made spam-complaint rate an existential metric. Every added touch is complaint surface now; a cadence that was merely annoying in 2019 can burn a domain today — see Deliverability and compliance essentials.
Default if the user has no preference: Moderate 3-5 — it collects the follow-up replies every large corpus agrees exist, without the complaint surface of a 15-touch cadence.
References
- Sam Nelson — Agoge Sequence (Substack, ~2018-19)
- Sam Nelson — Agoge Sequence writeup (guest article, Outreach blog, n.d.)
- 30MPC — Sam Nelson's Agoge Sequence (tactic writeup, n.d.)
- Belkins — Sales follow-up statistics (7.5M-email omnichannel study, 2026)
- Woodpecker — Cold email statistics (20M+ email corpus, updated Jun 2026)
- QuickMail — Cold email statistics (65M journeys and emails, Dec 2023)
- Corey Haines — marketingskills cold-email skill (open-source skill pack, MIT, 2025-26)
- LeadMagic — gtm-skills (open-source skill pack, MIT, 2025-26)