How to QA a campaign launch before it spends
The pre-flight checklist run by tools instead of eyeballs: settings, budgets, tracking, targeting and destinations verified before money moves.
IntentCatch the launch errors that cost real money in the first 48 hours, before the first impression.
- 5
- Steps
- 14
- Tools
- 4
- Rules
- 3
- Failure modes
When to use it
Use it when
- Every launch that matters.
- Especially launches created in a hurry, by a new team member, or cloned from an old campaign.
Do not use it when
- As a post-mortem: after spend has run, use campaign-anomaly-investigation.
What it needs first
Required
Preconditions
- The campaign exists, paused, on the ticked account.
- The launch brief available: intended budget, dates, market, destination.
Inputs
campaignrequired- The campaign to check.
briefrequired- Intended settings to verify against.
Procedure
In order. Every tool named is one the gateway ships, and links to its reference.
Read the campaign's actual settings: status, dates, budget object, bid strategy.
Compare every setting to the brief and list the diffs.
Verify conversion wiring live for the account before spend depends on it.
Check destinations respond, via the landing page view for existing URLs.
Deliver the go/no-go with the diff list; resuming the campaign is a Safe Write for a human to confirm.
Decision rules
The observable condition, what it lets you conclude, and what takes the conclusion back.
When
The budget object is shared with other campaigns and the brief assumed dedicated
Conclude
No-go until resolved: a shared budget makes the launch spend unpredictable by design.
When
The conversion action the bid strategy optimises to fired zero times in 7 days
Conclude
No-go: launching onto a dead conversion optimises toward nothing.
When
Dates, geo or language differ from the brief
Conclude
Diff listed, launch held: these three are the classic silent clone errors.
When
Everything matches
Conclude
Go, with the checklist as the record: a clean QA is a deliverable too.
Evidence every conclusion must carry
- Every check quotes the read value against the brief value.
- The go/no-go names the blocking diffs, or the absence of any.
Where the agent stops
- The campaign is resumed only through Safe Writes after human confirmation; the QA itself changes nothing.
- Campaigns the gateway creates are always created paused, which is what makes this checklist runnable at all.
How it goes wrong quietly
The cases where the analysis is wrong and still looks right. Read them before trusting a number.
- Cloned campaigns inherit invisible settings (exclusions, dayparting) nobody re-reads; the diff must cover the ones the brief never mentioned.
- Tracking verified account-wide can still miss the campaign's specific conversion action; check the one being optimised to.
- Timezone: midnight launches in the account timezone are not midnight in yours.
What the answer contains
- 01Diff list
- Read vs brief, setting by setting.
- 02Go / no-go
- The verdict with blocking items.
- 03Launch record
- The verified state at launch, for the post-mortem that hopefully never happens.