A cold email sequencer sends a planned series of messages from your mailboxes, stops the series when someone replies, and reports on what happened. That is genuinely useful and it is also the problem: automation scales whatever you give it. A sequence with a weak list and a four-step follow-up pattern becomes a complaint engine faster by hand than any human could manage.
The settings that decide your outcome are not the ones in the marketing material. They are the pacing, rotation and tracking options, and most tools ship them at defaults tuned for throughput rather than placement.
What the tool is responsible for, and what it is not
The sequencer's job: scheduling, personalisation merge, reply and bounce detection, stopping on reply, suppression lists, reporting.
Not the sequencer's job: your list quality, your warm-up, your DNS, or knowing which folder your mail landed in. Every sequencer reports "delivered" when the receiving server returned a 250, which says nothing about the inbox. A campaign in the spam folder reports as 100% delivered.
That gap is the single most important thing to understand about this category. If you optimise against sequencer metrics alone, you can watch a campaign look healthy for weeks while its reply rate quietly goes to zero.
The settings to fix before your first live send
| Setting | What to set it to | Why |
|---|---|---|
| Per-mailbox daily cap | Well below the platform ceiling, raised gradually | The ceiling is a limit, not a target |
| Sending window | Working hours in the recipient's time zone | A mailbox emitting mail at 03:00 local is a pattern |
| Delay between sends | Minutes with randomisation, not a fixed interval | Even spacing is machine-shaped |
| Mailbox rotation | Across several mailboxes and domains | Limits the blast radius when one gets filtered |
| Stop-on-reply | On, and across the whole sequence | Following up after a reply is the fastest route to a complaint |
| Bounce handling | Auto-suppress on hard bounce, pause sequence on a bounce spike | Bounce rate is your earliest warning |
| Link tracking | Off for cold outreach if you can | Each redirect adds a domain whose reputation is not yours |
| Open tracking | Off, or treated as unreliable | Apple Mail Privacy Protection and corporate proxies pre-fetch images |
| Thread follow-ups | Same thread, shorter each time | Gives the reader context and keeps the message short |
| Sequence length | Initial plus two follow-ups | The third and fourth bump is where complaints come from |
| Unsubscribe handling | Honour immediately, suppress globally | CAN-SPAM allows 10 business days; do it instantly anyway |
| Suppression list | Global, across all campaigns and mailboxes | Two campaigns reaching the same person is what gets reported |
Reply detection is worth testing yourself
Stop-on-reply is the feature that protects you most, and it fails in ways you will not notice from the dashboard. Test all of these before you trust it:
- An out-of-office auto-reply. Does it stop the sequence? It should *not* count as a real reply, but it also should not trigger the next step while the person is away.
- A reply from a different address at the same company. Does it stop the sequence for the original contact?
- A reply that arrives while the next step is already queued.
- A bounce message. Is it classified as a bounce or as a reply?
- A forwarded thread with a new participant.
Each of these has a sensible answer and tools differ. The expensive failure is a follow-up sent to someone who already said no.
Volume per mailbox is the number people get wrong
Sequencer capacity is warmed mailboxes multiplied by a sustainable daily rate, and the second number is much smaller than the platform limit. A new mailbox needs weeks of gradual warm-up before it carries anything, and it needs to stay warm afterwards because reputation is a rolling judgement about recent behaviour, not a badge you earn once.
The failure pattern is consistent: a team buys thirty mailboxes, configures the sequencer to its default cap, sees a week of apparently fine numbers, and then placement collapses across all thirty at once because they were all ramped identically on the same infrastructure in the same week.
Ramp per mailbox, stagger when mailboxes come online, and keep per-mailbox volume low enough that losing one does not matter.
Choosing between tools
The feature lists converge, so evaluate on these instead:
- Does it connect over OAuth to Gmail and Microsoft, rather than storing a password for basic SMTP? Microsoft has been retiring basic authentication.
- Can you turn link and open tracking off entirely, per campaign?
- Is the suppression list global across campaigns, mailboxes and teammates?
- Can you set randomised delays and real sending windows per mailbox time zone?
- Does it export cleanly? Contacts, sequences, reply history. Lock-in here is expensive.
- Does it pause automatically on a bounce or complaint threshold you define?
- What does it do on a 421 or 451 soft rejection — retry sensibly, or hammer the receiving server?
Treat current pricing and feature claims from any vendor in this category as unverified until you confirm them directly; the tools change quickly.
What not to do
Do not use a sequencer before warm-up. It is the most common way a new domain gets burned in week one.
Do not run your primary company domain through it. Use separate outreach domains so a bad sequence cannot touch your invoices and password resets.
Do not accept default caps. Defaults are set for throughput, and the cost of being wrong lands on you.
Do not add steps to fix a low reply rate. More follow-ups raise complaints, and Google asks bulk senders to keep reported spam under 0.3%. A weak reply rate is a targeting problem.
Do not let two campaigns share a contact. Global suppression, every time.
Do not report open rate as a result. It includes machines, and it will send you optimising the wrong thing.
Do not trust "delivered". It is an SMTP acceptance, not a folder.
Where MailPilot fits
MailPilot covers the two things a sequencer structurally cannot. It warms every mailbox before and during live sending — over OAuth on Gmail and Outlook, or over SMTP/IMAP for SendGrid, Amazon SES, Mailgun, Postmark and custom SMTP — and it measures real inbox placement per provider through a seed-inbox network, so you can see whether Gmail put the campaign in the inbox, in Promotions, or in spam. It also checks SPF, DKIM and DMARC per sending domain and tracks per-mailbox reputation daily, which is what tells you to pause a mailbox before the sequencer tells you your replies have stopped.
Ready to reach the inbox every time?
Automated email warmup across real peer mailboxes. Live inbox placement monitoring. Free 7-day trial - no credit card required.
Start free trial