Most people set up Smartlead correctly and then get the campaign wrong. They spend time on the domain registration, the DNS records, the warm-up schedule (all correct priorities; I covered the full infrastructure checklist in this earlier post), and then they build the sequence itself in a hurry: one email, three days between steps, ship it. The sequence is where campaigns actually live or die, and it is the part that gets the least deliberate attention.
This is the setup I run inside Smartlead for every campaign before anything sends. It takes roughly two hours the first time and about thirty minutes once you have templates. Those two hours have saved me more than a few burning domains.
Start with the inbox pool
Smartlead's core mechanism is rotating across multiple sending inboxes so no single address sends too many messages per day. The question is how many inboxes to use and how to organize them.
My rule: one inbox per 30 to 40 contacts per day, never more. If the campaign targets 300 people over 10 days, I want at least 3 inboxes in the rotation, ideally 4 or 5. Each inbox needs to be on a domain that has completed at least two to three weeks of warm-up and is sending at a stable reputation. I check the warm-up status in Smartlead's inbox health panel before including any inbox in a live campaign. An inbox that shows any red flags in the deliverability score gets pulled from the pool until it recovers.
The other rule I never break: campaign inboxes live on outreach-only domains, never the main business domain. The separation is not optional. A warm-up failure or a spam rate spike on an outreach domain is recoverable. The same thing on your primary domain takes your entire company's email credibility with it.
Building the sequence steps
A cold sequence needs four to five touchpoints. More than that and you are pestering; fewer and you have not given the campaign a real shot. Here is the cadence I use as a starting point, which I adjust per industry and list quality.
- Day 0: First email. Specific, relevant, short. Under 150 words. One clear observation or ask, not three.
- Day 3: First follow-up. One or two sentences. Acknowledge you have not heard back. Ask whether the first email landed at a bad time.
- Day 7: Second follow-up. A different angle or a resource. A little more context if the first message did not resonate.
- Day 14: Third follow-up. The break-up email. Short. "If this is not relevant, happy to stop." These often generate the highest reply rate because they create a natural close.
- Day 21: Optional. A final note only if you have something genuinely new to add, a product update, a case study, a referral. Skip this step if you are just filling the cadence.
Inside Smartlead, each step gets its own subject line option. For the first step I always set up at least two subject line variants for A/B testing and check which outperforms after the first 50 sends. That winner carries into the next campaign targeting a similar audience.
Personalization that actually earns its place
Smartlead supports variable substitution in the copy: {{first_name}}, {{company}}, and any custom columns you define in the upload CSV. The mistake most people make is treating personalization as a checkbox. Add {{first_name}}, feel like it is personal, move on.
What consistently outperforms is a custom variable for one specific, verifiable thing that makes this person worth reaching out to. I call that column {{trigger}} in my CSV. It might be a recent company announcement, a job change I saw on LinkedIn, a tool they are publicly known to use, or an industry development that affects their role. One line that proves I did thirty seconds of research beats three generic lines of "I saw you work in X" every time.
For campaigns targeting companies with 50 or more employees, I often add a {{company_size_tier}} variable: "growing teams," "mid-size orgs," "larger organizations." That lets me vary one sentence of the pitch slightly without maintaining three entirely separate sequences. Small change, noticeable lift in reply rates because the framing feels less like a mass blast.
Setting up reply classification
Smartlead classifies incoming replies automatically: Interested, Not Interested, Auto-reply, Wrong Person, Information Requested, Meeting Booked. The model does a reasonable job, but it needs to be checked, not just trusted.
Before any campaign launches, I verify three things in the reply classification settings.
First: "Not Interested" is wired to the global suppression list. When someone opts out, they come off the sequence immediately and never receive another email from any campaign running in the same workspace. Smartlead handles this automatically, but I confirm the suppression sync is active, because a misconfigured workspace can break it silently.
Second: "Interested" and "Meeting Booked" both have a notification set up. These categories need a human response, often the same business day. I route them to a Slack notification or an email alert so they do not get buried in the unified inbox while the campaign runs in the background.
Third: "Wrong Person" triggers a pause on that contact. Someone telling me I have the wrong address is not someone to continue sequencing.
The signals I watch in the first 48 hours
Once a campaign goes live, the first two days tell you almost everything you need to know about whether it will run cleanly.
Open rate: Should be above roughly 40% if subject lines and inbox rotation are healthy. Below 30% after the first 50 to 100 sends usually indicates a deliverability problem, not a copy problem. Fix the infrastructure before rewriting the message.
Reply rate: Above 3% is a reasonable benchmark for a cold campaign with solid list quality and targeting. Below 1% after 100 sends is a signal that the message itself needs work, not necessarily that the list is bad.
Bounce rate: Stays below 2% on a clean list. Above 5%, I pause the campaign immediately and audit where the contacts came from. High bounce rates damage sending reputation faster than almost anything else.
Spam folder rate: I run Smartlead's inbox placement checker before launch and again after the first 50 sends. If anything starts landing in spam, the campaign stops. Scaling a campaign with a deliverability problem destroys the domain's reputation, and rebuilding that takes weeks, not days.
If anything looks wrong in the first 48 hours, I pause before scaling. It costs two days. The alternative costs a domain.
The one ceiling I do not cross
Forty emails per inbox per day. That is the upper limit for any inbox in a campaign rotation, even if Smartlead lets you set it higher. Staying under 40 gives buffer for warm-up interactions, test sends, and spikes in incoming replies. The sending pools I have seen burn the fastest are always ones where someone pushed an inbox to 80 or 100 per day because the campaign needed to move faster. The campaign moved faster. The domain did not survive it.
If you need higher volume, the answer is more inboxes with proper warm-up, not more sends per inbox. Patience in setup is not a constraint. It is what makes the campaign actually run.
Sources
- Smartlead.ai: inbox rotation, sequence builder, reply classification documentation
- Five Steps I Take Before Any Cold Email Campaign Goes Live: domain registration, DNS setup, and warm-up prerequisites
Running outbound and want to compare notes on a sequence that is not performing the way the deliverability metrics suggest it should? Drop me a line.
Comments