Outbound Campaign: Lead Time First

Why outbound launches slip

An outbound launch almost never slips because too little work got done. It slips because two clocks are running and only one of them is in the plan. The first measures effort and gets shorter when you add people. The second measures time, and nobody shortens that one.

Every team knows the first clock. Sharpen the target segment, build the list, write the approach, agree the sequence. That is the part you can talk about in a kickoff, and it is the part that feels like progress, because you watch it grow every day.

The second clock runs quietly. It starts the moment the first mailboxes exist, and it keeps running whether the team is working on the list or on holiday. Start it after the kickoff and you have started it exactly when everything else was already finished.

Our position: we stand the sending infrastructure up on day one of an engagement rather than after the kickoff, because it is the one variable in the project that effort does not move. Everything else can be parallelised or reinforced. This cannot.

Warm-up is a calendar cost, not a workload

A freshly created mailbox on a new domain has no sending history. Nobody can supply that history; it only comes from sending and receiving in small volume over a stretch of time. Start at full volume and you will not see an error message. You will see it later, in reply rates nobody can explain.

The mechanics are publicly documented, and by the receiving side itself. Google's sender guidelines say: "Start with a low sending volume to engaged users, and slowly increase the volume over time", and warn explicitly against jumping the volume — "Avoid introducing sudden volume spikes if you do not have a history of sending large volumes."

Google does not say how long that takes. A sending service does: by Twilio SendGrid's own documentation, its automated warm-up schedule ends on day 41. That is not an estimate from a blog post. It is the schedule the tool itself runs.

Our position: we plan the lead time in weeks rather than days, because anything tighter is planned against the documentation of your own tool. And we run the warm-up in parallel with targeting and copy, not after it. No campaign mail goes out during that window. That is not caution — it is the entire purpose of the window.

Two decisions that get expensive after setup

Who writes under which name, and from which address, often gets decided in passing. Neither is cheap to change once the mailboxes exist. A change means new mailboxes, and new mailboxes mean paying the lead time a second time. So both belong in front of the clock, not inside its run.

The senders. Who appears as the sender moves reply rates more than any phrasing does. A founder or a head of department gets read more often than a role the recipient cannot place. Those names are locked once the mailboxes are provisioned, so two things get settled first: the exact spelling of every name, and a named stand-in for each sender in case that person leaves.

The domains. Outbound does not run on the main domain; it runs on dedicated sending domains. The common mistake is not the choice but its order. A domain gets bought, configured and warmed up, and only once deliverability stays poor does somebody check whether it already sits on a blocklist. A second-hand domain carries a history nobody can undo.

Our position: we check domains against the relevant blocklists before selection, and we settle the selection in writing rather than in a meeting. Not out of formality, but because both decisions need a date that falls before the clock starts. Hunting for domain names live in a call burns an hour and rarely produces a result anyone defends afterwards.

Deep Dive Guide · 17 pages · free

From signal to sales pipeline

Whether your market carries volume, which three to five signals count and where the chain would break. Plus the checks before the first send.

Explore guide →

The link that sits on somebody else's address

When an email measures whether somebody clicked, that click travels through an intermediate address. By default the address belongs to the sending tool and is shared across all of its customers. Its reputation is therefore built by senders you neither see nor influence. It still lands on your emails.

That is not negligence on the vendors' part. It is a sensible default for getting started, and the wrong one the moment predictable pipeline depends on this infrastructure. In Lemlist, the tool we run sequences in, the custom tracking domain lives under the sending settings; other sequencers name the menu differently, but the record you add at the domain provider is identical.

Technically it is a pointer record in the domain's naming system (a CNAME), and it is one per sending domain, not one per mailbox. Three failure modes here are silent, with no error anywhere:

1. The value is entered without its trailing dot. Some providers then append your own domain, and the record resolves to nothing.
2. The wrong record type is chosen. The picker offers several; only CNAME is correct.
3. The value is typed rather than pasted, for instance from a video or a screenshot. One wrong character is enough.

Our position: the acceptance test is the green status in the sequencer, never the saved form. The distinction sounds pedantic and is the whole point. A saved form proves somebody clicked save. A green check in Lemlist proves the record actually resolves on the public internet. Until then, no campaign runs with link tracking on.

What that means for the quarterly plan, read backwards

A launch date estimated from the amount of work left is not a commitment. It becomes one only when you calculate it backwards. Start at the first send you want and subtract the warm-up window. What sits before that are the two approvals and the domain check. Everything else runs alongside.

The difference is not academic. Every week lost at the start is a week of missing pipeline at the end, and because a full sales cycle sits between the first send and a signed contract, it only becomes visible a quarter later. By then nobody is looking for the cause in a DNS record from March.

How you can tell the order is working

The proof is not that somebody works faster. It is that a team with no experience in active outreach builds a running acquisition motion within months. That happens when the lead time sits at the beginning rather than at the end. Two published examples show the order of magnitude.

The starting point at knk is recorded verbatim in the case study: "No outbound process: Before working with SalesPlaybook, they had no outbound sales strategy and only emailed existing subscribers." From there to a running, automated pipeline took less than six months. The infrastructure lead time fits visibly inside that window when it sits at the front. Put it at the back and it extends the window by exactly its own length.

Where the replies go, before the first one arrives

One more item belongs before the launch, even though it does not look like infrastructure. Who answers on the customer side is the most common breaking point in outbound, and not because the approach was weak. It is because nobody feels unambiguously responsible, and you find that out when the first interested reply sits untouched for three days.

Two models work, and both only work if you pick one. Either a named person takes every reply. Or a workflow in HubSpot routes them automatically by an attribute such as region or segment. What does not work is a shared address everyone may look at and nobody has to.

Our position: every positive reply gets a named owner in HubSpot before the first campaign goes out. The reason is the same as for the lead time. You can set this up afterwards, but the replies that went cold in the meantime do not come back.