What to check before you launch a campaign

The answer, before the send button

An outbound campaign checklist should confirm the audience, sequence structure, message composition, thread behavior, testing setup, suppression rules, send projection, and prospect movement. The useful outbound campaign checklist is less about catching typos and more about forcing the campaign to fail in a test environment first.

That sounds obvious. Teams still launch campaigns where follow-ups open new threads, replied prospects remain active, and a test promotes a winner after a handful of sends.

I've seen this happen because everyone reviews the copy and nobody reviews the machinery around it. The email sounds fine. The campaign is still broken.

Start with the campaign you actually built

First, work out whether the campaign is tiered or flat. Don't assume the structure from the copy.

A tiered campaign gives different groups their own sequences and message logic. A flat campaign sends the same basic path to everyone. Either can work. Problems start when the campaign is built one way and judged as if it were the other.

Say a 40-person cybersecurity company is targeting 600 prospects after hiring its first enterprise sales rep. The list includes security directors at larger companies and hands-on IT managers at smaller ones. If both groups enter one sequence, the team may blame the copy when the real problem is that the campaign has two different buying situations hiding inside one report.

For a tiered campaign, check each tier's audience, sequence, timing, and offer. A senior buying group shouldn't inherit the sequence written for an operational user just because both groups were uploaded together.

For a flat campaign, ask whether the path is genuinely common. If timing, pain, or the next action changes by audience, a flat structure will make the results hard to interpret. You'll know something underperformed, but not what.

That's the first opinion I'd push on most teams: they over-focus on message variants when the campaign structure is already muddy. A cleaner sequence usually teaches you more than another round of copy edits.

Look at composition, not just wording

Open every step and check how the message is assembled. The system supports four composition methods, from fully personalized to templated with a written opener. The setting for each step should match the reason the prospect is there.

A fully personalized message needs enough source material to produce a real observation. A templated message with a written opener needs an opener that adds something, not one that restates the template with a prospect's company name dropped in.

A variable isn't personalization. A first name isn't relevance. "Saw that Acme is hiring" only matters if the campaign explains why that hiring is a useful reason to contact Acme now.

Here's a practical test. Take a campaign aimed at 100 recently funded software companies. Pick three prospects from different company sizes and industries. Read the opener without looking at the template. Can you tell what changed for each person? If not, the campaign probably isn't using its composition method properly.

Check where the individual detail enters, what stays fixed, and whether the fixed section still makes sense for the whole audience. If the method can't create the intended distinction, rewriting the sentence won't fix the setup.

Send one prospect through the whole thing

Create a test prospect and receive the campaign as if you were the buyer. Don't stop at the preview.

The first message, every follow-up, and every reply branch need to behave like one conversation. Follow-ups should use the original reply headers and provider thread reference. The recipient should see one thread, not five cold emails with matching subject lines.

This is where preview screens lie by omission. A message can look perfect on its own and still create a new thread once it's sent. The result is duplicate subject lines, broken context, and a sender who looks careless even though the copy was reviewed three times.

Reply to the first message from the test inbox. Then inspect what happens next. Does the follow-up stay in the same thread? Does the reply branch use the right message relationship? Does the campaign stop or change path when it should?

Changing the subject line isn't the only way to break threading. A familiar subject doesn't prove that the messages belong to the same conversation.

Test the decision, not the feature

A/B testing works at sequence level and at individual step level. Attach the test to the decision you want to make.

If you want to know whether one complete sequence works better than another, test at sequence level. If you're comparing two openers, test the relevant step. Don't test an entire path when the question is about one sentence. You'll get a result, but you won't know what caused it.

For example, a 12-person sales team at a recruiting software company might test whether a problem-led opener beats a referral-led opener for founders. That belongs at the opening step. Testing two entirely different sequences would also change the timing, follow-ups, and calls to action. The result would answer a much larger question than the team intended.

Check the variants, the step where each applies, and the conditions for automatic winner promotion. Early movement is not a verdict. If the system promotes a winner before the sample is useful, the campaign can lock onto random noise and make the report look more certain than it is.

The opposite mistake happens too: teams change the opening, offer, timing, and audience at once, then call the result an A/B test. That's not a test. It's a new campaign.

Suppression comes before volume

Review who must be excluded before looking at how many people will receive the campaign.

Suppression should remove prospects who are already excluded by policy or who shouldn't receive this particular campaign. That might include current customers, open opportunities, recent responders, employees, competitors, or people already contacted by another team.

The position of the rule matters. A clean audience file can still produce a bad send if exclusions are applied after the audience has been assembled.

Then inspect the 30-day send projection. The first-day count isn't the campaign. Follow-ups accumulate. Replies create branches. A launch that appears to send 80 messages on day one may create a much larger workload over the next month.

Imagine a 25-person agency launching to 1,200 contacts with four follow-ups. The team has enough capacity for the initial send, but not for the reply review and manual handoffs created by week three. The projection should expose that before launch, not after someone's inbox becomes the bottleneck.

Decide what a reply does

Every prospect has a pipeline position. Replies may move that prospect automatically, and each campaign has a setting to turn that behavior off.

Decide this before launch. Don't leave it as a default.

If automatic movement is on, send a test reply and confirm that the prospect advances to the intended position. If it's off, confirm that the campaign doesn't quietly depend on a rep remembering to update it.

This is a small setting with an annoying failure mode. A reply can arrive, but the prospect still appears active in the outbound campaign. Someone then sends the next follow-up to a person who already answered.

The reverse can happen too. Automatic movement may be wrong for a campaign where every reply needs manual review before the prospect changes position. A short "not now" and a qualified buying conversation shouldn't necessarily trigger the same transition.

Use this last pass before launch

Run the campaign as both a recipient and an operator. Check the points that are easiest to miss:

Audience and suppression: Is everyone included for a reason, and are current customers, open opportunities, recent responders, and other exclusions removed?

Campaign shape: Is the campaign tiered or flat, and does the sequence match that choice?

Composition: Does each step use the intended method, with enough real information to support the level of personalization?

Thread behavior: Do follow-ups stay in the original conversation using the correct reply headers and provider thread reference?

Testing: Is the test attached to the sequence or step that matches the actual question?

Send projection: Does the 30-day view include follow-ups and the work created by replies?

Pipeline movement: Will replies move prospects automatically, or is manual review intentional?

Launch state: Has someone tested the complete path from the inbox and the campaign operator's view?

The point isn't to approve a campaign because every box has a checkmark. It's to make the likely failure happen while changing the setup is still cheap.