Walkthrough-First Outbound, Run Through Reply.io
Most agency cold email tries to close the argument inside the email — attach the case study, list the problems, link the deck. But the moment a web designer is actually persuasive is the moment they're sharing their screen: pulling up the prospect's site, dragging the viewport down to phone width, and showing — live — what the fix looks like. PDFs get skimmed; a 15-minute screen-share gets remembered. The highest-leverage CTA a design agency can run isn't "can I send you our portfolio?" It's "give me 15 minutes and I'll walk you through it on your own site."
Reply.io is built for exactly that kind of pursuit, because booking a live walkthrough is rarely a one-channel job. Its multichannel sequences put email, call tasks, and LinkedIn steps in one flow — the email plants a single observation, the call task is your moment to offer the walkthrough voice-to-voice, and the LinkedIn step keeps your work scrolling past the prospect in between. What Reply.io doesn't do is provision the sending infrastructure underneath: the secondary domains, mailboxes, and dedicated IPs the email channel runs on. That's where ColdRelay fits. This guide covers how design agencies wire the two together, and how to structure sequences where every step exists to book the screen-share.
Why Run Reply.io on ColdRelay Infrastructure
Reply.io is a sales engagement platform — it orchestrates email steps, call tasks, and LinkedIn touches across a sequence, manages contacts, and pools responses into a unified inbox. It sends email from whatever mailboxes you connect to it; it doesn't provision domains, create mailboxes, or own the sending reputation underneath. That layer is yours to supply.
The walkthrough model makes that layer matter in a specific way: in a multichannel sequence, the email step is the setup for everything that follows. The call task script assumes the prospect saw your observation; the LinkedIn connection request reads naturally because your name arrived in their inbox two days earlier. If the email lands in spam, you're not just losing a send — you're cold-calling cold and connecting with a stranger, and the whole choreography collapses into three disconnected interruptions. ColdRelay keeps the first domino standing: mailboxes provision on isolated Azure tenants with dedicated IPs, with SPF, DKIM, and DMARC pre-configured, ready in about an hour, holding 95%+ inbox placement. No warmup waiting period either — warmup runs continuously as part of each mailbox's 4 sends/day budget (2 outbound + 2 warmup), so the sequence can go live the day the mailboxes do.
The pairing is additive, not competitive: ColdRelay is the infrastructure, Reply.io is the engagement layer on top. You keep Reply.io's multichannel sequences, Jason AI, and unified inbox — you just run the email channel on mailboxes built to land.
Visit Reply.io →Connecting ColdRelay Mailboxes to Reply.io
Provision mailboxes on ColdRelay
Pick secondary domains adjacent to your agency brand — never the domain that hosts your portfolio. The walkthrough model is calendar-constrained, not send-constrained: you can only deliver so many live screen-shares a week, so most agencies run 15-30 mailboxes. ColdRelay supports 100-150 mailboxes per domain, so the whole pool fits on a single domain. Everything provisions on isolated Azure tenants with dedicated IPs in about an hour, with SPF, DKIM, and DMARC already configured.
Connect the mailboxes as email accounts in Reply.io
Export your mailbox credentials from the ColdRelay dashboard, then connect each mailbox in Reply.io via SMTP/IMAP under your email account settings. Once connected, Reply.io can distribute sequence emails across the pool, and replies from every mailbox flow into the unified inbox alongside your call outcomes and LinkedIn responses — one queue for the whole pursuit.
Set per-mailbox sending limits to match the ColdRelay budget
Use Reply.io's per-mailbox sending limits to cap each account at 2 outbound emails per day — mirroring ColdRelay's per-mailbox budget of 4 sends/day total, split 2 outbound + 2 warmup. ColdRelay's warmup runs continuously in the background, so Reply.io only ever handles the outbound half; don't layer additional warmup on top.
Build the multichannel sequence around the walkthrough CTA
Create a Reply.io sequence that interleaves the three channels toward one ask. Step one: an email with a single specific observation about the prospect's site — one, not a list — and the offer to show the fix on a 15-minute screen-share. Step two: a LinkedIn profile visit and connection request, so your face and your pinned portfolio work attach to the name in their inbox. Step three: a call task with a script that opens from the observation ("I emailed you about the checkout flow on mobile — I can show you the fix in 15 minutes, screen-share, no deck"). Step four: a short breakup email that restates the walkthrough offer and nothing else. Load your prospect list into Reply.io's contact management with a custom field for each prospect's observation, so every step can reference it.
Turn on Jason AI for triage and launch
Enable Jason AI to assist with reply handling: let it categorize responses (interested, objection, not now, wrong person) and draft answers to the routine ones, while you personally handle anything that smells like a yes — a walkthrough booking deserves a human confirming the time and asking one scoping question. Launch the sequence, and treat the unified inbox as your booking desk: the metric that matters is walkthroughs on the calendar, not opens.
The Walkthrough-First Reply.io Playbook
Sell the 15 minutes, not the redesign
The email's only job is to make the screen-share feel like a small, safe yes. One observation, one sentence of why it costs them money, one ask: 15 minutes, their site on your screen, you show the fix live. No attachments, no portfolio PDF, no "capabilities overview" — anything you attach gives the prospect a way to feel finished without meeting you. The walkthrough is where a designer's actual advantage lives: watching someone fix your site in real time is a hiring decision forming in the prospect's head, and no attachment does that.
Hold the evidence back for the call task
In a single-channel sequence, you'd spend all your proof in the email. In Reply.io's multichannel flow, ration it: the email names the problem, but the demonstration belongs to the live moment. When the call task comes up, your script has somewhere to go — "I can describe the fix, but it's faster to show you" — and the call's purpose is never to sell the project, only to calendar the walkthrough. A prospect who won't take a pitch call will often take a call whose stated agenda is scheduling something useful and short.
Use the LinkedIn step as a portfolio that scrolls past them
Your LinkedIn presence is the one channel where showing work isn't an imposition. Before the connection step fires, make sure the profile it lands on is doing its job: a pinned before/after, a recent post walking through a real fix, a headline that says what you build. Then let Reply.io's LinkedIn steps run the touch — a profile visit before the first email, a connection request after it. Between email and call, the prospect who checks you out finds a working designer's feed, not an empty avatar — and every post you publish keeps your work drifting past everyone who accepted but hasn't booked yet.
Run the walkthrough as a give, and let the close ask itself
Deliver the 15 minutes as promised: their site on screen, the observation from the email demonstrated live, the fix mocked up or sketched in real time, one or two adjacent issues noticed out loud but not belabored. Don't pivot to a pitch at minute twelve. End with "want me to write up what a proper pass at this would look like?" — the prospect has just watched you work, which is more vetting than most agencies ever get, and the proposal request usually comes from their side of the call. Walkthroughs that don't convert still compound: you're the designer who showed up, shared a screen, and fixed something for free.
Typical Walkthrough-First Outbound Benchmarks (Reply.io + ColdRelay)
| Metric | Benchmark | Notes |
|---|---|---|
| Inbox placement rate | 95%+ | Dedicated IPs and isolated tenants — the email step is the setup for every later channel, so it has to land |
| Reply rate across the full sequence | 4-8% | Multichannel pursuit outperforms email-only; the call task converts prospects who read but never reply |
| Positive replies that book a walkthrough | 40-60% | A 15-minute screen-share is a low-friction yes compared to a discovery call or a proposal request |
| Walkthroughs that turn into proposals | 25-40% | Prospects who watch the fix demoed live on their own site rarely need further convincing to see a scope |
| Outbound capacity per mailbox | 2/day | 4 sends/day total per mailbox — 2 outbound + 2 warmup; calendar capacity, not send capacity, is the real ceiling |
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, IPs, and isolated Azure tenants are included — and because walkthrough capacity is bounded by your calendar rather than your send volume, most agencies run this model on a small, steady pool.
Reply.io is billed separately on its own subscription for multichannel sequences, call tasks, LinkedIn steps, Jason AI, contact management, and the unified inbox — priced per its current plans.
The two bills buy the two halves of the pursuit: ColdRelay's per-mailbox infrastructure makes sure the opening observation reaches the inbox, and Reply.io's subscription orchestrates everything that happens after it — the LinkedIn touch, the call task, the booking. Against the project value of a single signed redesign, the stack is a rounding error per walkthrough booked.
| 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. Reply.io handles the engagement side — multichannel sequences, call tasks, LinkedIn steps, Jason AI reply handling, contact management, and the unified inbox. ColdRelay provides the underlying domains, mailboxes, and dedicated IPs that Reply.io's email steps send from. You use them together: infrastructure underneath, engagement orchestration on top.
Why book a screen-share instead of just attaching the portfolio or an audit PDF?
Because an attachment lets the prospect feel finished without ever meeting you — and it flattens your strongest asset into a file. A live walkthrough shows the prospect their own site being improved in real time, which is the closest thing to a work trial a cold prospect will ever give you. The email's job is to earn 15 minutes, not to win the argument; everything you'd have put in the PDF performs better on a shared screen with the prospect asking questions.
Do the call tasks and LinkedIn steps affect the ColdRelay sending budget?
No — only email steps draw on the mailbox budget. Each ColdRelay mailbox sends 4 emails/day total (2 outbound + 2 warmup), and Reply.io's per-mailbox sending limits keep the outbound half inside that line. Call tasks and LinkedIn steps run on their own channels entirely, which is part of why the multichannel model suits a conservative send budget: two of your touches per prospect cost zero sends.
Should I let Jason AI handle replies that want to book a walkthrough?
Use it for triage, not for the booking itself. Jason AI is good at categorizing responses and drafting answers to routine objections and not-nows, which keeps the unified inbox manageable across a multichannel sequence. But a reply that's leaning yes is the most valuable object your outbound produces — confirm the time personally, ask one scoping question ("anything specific you want me to look at?"), and arrive at the walkthrough already prepared. Automation earns the conversation; the designer should be the one who shows up to it.