Compound-Touch Agency Outbound, Run Through Reply.io
Ask an agency owner where their last three retainers came from and the honest answer is rarely 'an email.' It's 'they'd seen us around' — an email they half-read, a name they recognized from LinkedIn, and then a call that suddenly wasn't cold anymore. Pitch meetings don't come from one great touch; they come from compound touches, where each channel makes the next one warmer.
Reply.io is built around exactly that mechanic: multichannel sequences put email steps, LinkedIn steps, and call tasks in a single flow with conditions between them, so the case-study email, the CMO-visible LinkedIn engagement, and the call all fire in order — per prospect, automatically. ColdRelay is the layer underneath: the secondary domains, mailboxes, and dedicated IPs the email steps actually send from. This guide covers wiring the two together, and using Reply.io's sequence analytics to learn which channel books the meeting for each vertical your agency targets — because the answer is different for e-commerce founders than for B2B CMOs, and most agencies never find out.
Why Run Reply.io on ColdRelay Infrastructure
Reply.io's strength is orchestration: one sequence can open with a case-study email, follow with a LinkedIn step a couple of days later, and surface a call task for your team once the prospect has seen you twice. Jason AI helps draft steps and triage replies, the unified inbox keeps every conversation in one queue, and per-step analytics show where in the chain prospects actually convert. What Reply.io doesn't do is provision the email channel itself — it sends through whatever mailboxes you connect, and the domains, DNS, and IP reputation behind them are yours to solve.
That matters more in a compound-touch motion than in a pure email one, because email is the load-bearing first touch. If the case-study email lands in spam, the LinkedIn step that follows isn't a familiar name showing up — it's a stranger, and the call task at the end of the sequence goes back to being a cold call. ColdRelay keeps the first touch landing: mailboxes on isolated Azure tenants with dedicated IPs, SPF, DKIM, and DMARC pre-configured, live in about an hour, with 100-150 mailboxes supported per domain.
The pairing is additive, not competitive: ColdRelay is the infrastructure layer, Reply.io is the sequencing and orchestration layer on top. You keep Reply.io's multichannel steps, Jason AI, and unified inbox — you just anchor the whole chain on email that actually arrives.
Visit Reply.io →Connecting ColdRelay Mailboxes to Reply.io
Provision the email anchor on ColdRelay
Order mailboxes on one or two secondary domains close to your agency brand — ColdRelay supports 100-150 mailboxes per domain, and everything provisions on isolated Azure tenants with dedicated IPs in about an hour, with SPF, DKIM, and DMARC pre-configured. Compound-touch sequences run on tighter lists than spray campaigns, since every email prospect also becomes a LinkedIn touch and potentially a call task for a human.
Connect the mailboxes and set per-mailbox limits in Reply.io
Add each ColdRelay mailbox in Reply.io under Settings → Email Accounts via SMTP/IMAP, and use Reply.io's per-mailbox sending limits to cap each account at 2 outbound emails per day — matching ColdRelay's budget of 4 sends/day total per mailbox, split 2 outbound + 2 warmup. Warmup runs continuously on ColdRelay's side, so Reply.io only ever spends the outbound half and there's no waiting period before the first sequence.
Build one multichannel sequence per vertical
In Reply.io, create a separate sequence for each vertical your agency targets — e-commerce, B2B SaaS, healthcare, whatever your book looks like — using the same step skeleton: case-study email on day one, a LinkedIn step (profile view, then connection or post engagement the decision-maker will see) on day three, a second email referencing the first, then a call task. Separate sequences per vertical is what makes the channel analytics readable later; a merged list averages away the answer.
Load contacts with vertical and proof-point fields
Import prospects into Reply.io's contact management with custom fields for vertical and the matching case study, so the day-one email pulls a same-vertical proof point automatically — an e-commerce founder gets the e-commerce number, not your SaaS win. Let Jason AI draft step variants from your base copy, but keep the proof point human-chosen; it's the one variable doing the persuading.
Route replies to the unified inbox and work the call tasks same-day
Every email and LinkedIn reply lands in Reply.io's unified inbox — triage it daily, with Jason AI handling the first-pass sort of interested versus not-now. Call tasks deserve the same SLA: the task exists because the prospect has now seen your agency twice, and that familiarity decays in days. A call task worked the day it surfaces is a warm call; the same task a week later is cold again.
The Agency Compound-Touch Playbook for Reply.io
Sequence the touches so each one makes the next warmer
Order matters more than copy. The case-study email goes first because it plants a concrete number next to your agency's name. The LinkedIn step goes second because its only job is recognition — when the CMO sees your connection request or your comment on their post, the name should already ring a bell from the inbox. The call goes last because by then it isn't an interruption, it's a follow-up. Reply.io's sequence conditions enforce the order per prospect automatically; the mistake to avoid is running the three channels as three parallel campaigns, where the prospect meets you three times as three different strangers.
Make the LinkedIn step something the decision-maker actually sees
A profile view from an unknown agency does nothing. The LinkedIn steps that compound are visible ones: a substantive comment on the CMO's own post, a connection request that references the email ('sent over the CAC numbers from a brand in your space — this is me putting a face to it'). Reply.io schedules the step; the content of it is where the agency's judgment goes. The test: would the prospect plausibly mention 'oh, I saw you on LinkedIn' on the eventual call? If not, the step is invisible and the chain has a gap in it.
Read the per-step analytics by vertical, not in aggregate
This is the payoff of one-sequence-per-vertical: Reply.io's analytics show which step generates the booking for each audience, and the pattern is rarely the same twice. E-commerce founders often reply straight off the case-study email — the number does the work. B2B CMOs frequently go quiet on email and convert on the call task, after the LinkedIn touch made the name familiar. Once you can see that, you rebalance: heavier email investment in the verticals that book from the inbox, earlier and better-prepped call tasks in the ones that book on the phone. Agencies that read analytics in aggregate keep funding the wrong channel for half their list.
Size the list to your call capacity, not your send capacity
In a compound-touch motion, the constraint isn't email throughput — at 2 outbound sends/day per mailbox (of the 4/day total, with 2 warmup), even a modest ColdRelay pool outpaces what matters downstream: every prospect who advances becomes a call task a human has to work. If your team can make fifteen quality calls a day, a list that generates forty daily call tasks isn't ambitious, it's a backlog of decaying familiarity. Start with the call capacity, work backwards to list size, and let Reply.io's task queue — not your mailbox count — set the pace.
Typical Compound-Touch Agency Benchmarks (Reply.io + ColdRelay)
| Metric | Benchmark | Notes |
|---|---|---|
| Inbox placement rate | 95%+ | Dedicated IPs and isolated tenants outperform shared Google/Microsoft pools |
| Meeting rate (full email + LinkedIn + call sequence) | 1.5-3x email-only | Compound touches lift bookings most where email alone goes quiet; varies by vertical |
| LinkedIn connection accept rate after the case-study email | 30-45% | Roughly double a cold request — the email makes the name familiar before the request arrives |
| Call task to conversation rate (worked same-day) | 15-25% | Two prior touches make it a warm call; decays sharply if tasks sit more than a few days |
| Outbound capacity per mailbox | 2/day | 4 sends/day total per mailbox — 2 outbound + 2 warmup |
What It Costs: Reply.io + ColdRelay
You pay per mailbox per month for the infrastructure, with volume tiers that drop as you scale (see the table below). DNS, dedicated IPs, and isolated Azure tenants are included — and because compound-touch lists are sized to call capacity rather than raw volume, the pool is usually a modest, stable line item.
Reply.io is billed separately on its own subscription for multichannel sequences, Jason AI, the unified inbox, contact management, and analytics — priced per its current plans.
Infrastructure cost scales with mailbox count; Reply.io's cost scales with seats and plan tier. The expensive input in this motion is your team's call time — which is exactly what the per-vertical analytics protect, by telling you which prospects deserve it.
| Mailboxes | ColdRelay price / mailbox / month |
|---|---|
| 1–199 | $1.00 |
| 200–999 | $0.85 |
| 1,000–4,999 | $0.70 |
| 5,000+ | $0.55 |
Each mailbox sends 4 emails per day — 2 outbound to prospects + 2 warmup. ColdRelay provisions mailboxes on isolated Azure tenants with dedicated IPs; Reply.io handles the sending, sequencing, and inbox rotation on top.
Frequently Asked Questions
Does ColdRelay replace Reply.io?
No — they're complementary layers, not competitors. Reply.io is the sequencing and orchestration layer: multichannel sequences across email, LinkedIn steps, and call tasks, plus Jason AI, the unified inbox, and per-step analytics. ColdRelay is the infrastructure layer underneath: the secondary domains, mailboxes, and dedicated IPs Reply.io's email steps send from. You use both together — Reply.io runs the chain of touches, ColdRelay makes sure the email links in that chain actually land.
If the sequence is multichannel anyway, does email deliverability still matter that much?
More, not less — email is the anchor touch the rest of the chain leans on. The LinkedIn step works because the prospect recognizes your name from the inbox; the call task works because two prior touches made it warm. If the case-study email lands in spam, every downstream step degrades back to cold. That's why the email channel runs on ColdRelay infrastructure — isolated Azure tenants with dedicated IPs, SPF, DKIM, and DMARC pre-configured, 95%+ placement — so the first domino actually falls.
How do we find out which channel books meetings for each vertical we target?
Build one Reply.io sequence per vertical with the same step skeleton, then read the per-step analytics separately for each. Within a quarter the pattern is usually visible: some verticals reply directly to the case-study email, others go quiet on email and convert on the call task after the LinkedIn touch. The structural requirement is keeping verticals in separate sequences from day one — merge them into one list and the analytics average into mush, and you'll keep over-investing in the wrong channel for half your prospects.
How many mailboxes does this motion need, and is there a warmup wait?
No wait — ColdRelay mailboxes warm continuously as part of each mailbox's 4 sends/day budget (2 outbound + 2 warmup), so sequences can start the day the pool provisions, about an hour after ordering. Pool size follows call capacity rather than send ambition: a team that can work 10-15 call tasks a day is typically well served by 15-25 mailboxes (30-50 outbound sends/day), and with 100-150 mailboxes supported per domain, scaling later doesn't require new domains.