Testing programmes rarely die from bad ideas. They die from tests that were never powered, never documented, and never allowed to change anything downstream.
## Intake and prioritisation
Every hypothesis states the friction, the audience, the change and the metric it should move. Prioritise on expected value: traffic reaching the page, plausible lift, and the cost of building it. A ranked backlog kills the loudest-voice problem faster than any process document.
## Sizing and stopping rules
Decide the sample and the duration before launch, and write the stopping rule down. Most paid landing tests need whole weeks, not days, because weekday and weekend intent differ. If you cannot reach the sample in four weeks, the test is not worth running yet.
## Readouts that change spend
A readout is one page: what changed, what moved, what did not, and the recommended action for the media buyer. Losses matter as much as wins, because the audience-level reason a variant lost usually tells you where to point budget next.
Where this applies
- CRO
- PPC
Work with Northbeam Studio
This playbook comes from the work Northbeam Studio does. See what they offer, the results they can evidence, and how to start a conversation.
More from Northbeam Studio
- GuideFinding wasted spend in a neglected search accountA two-hour pass that reliably finds 15–30% of recoverable spend in accounts nobody has touched for a while, in the order we work through it.
- GuideSite architecture that search engines can actually followMost sites that plateau in organic search do not have a content problem. They have an architecture problem: too many shallow pages, no clear hub for any topic, and internal links that follow the navigation instead of the demand.
- DocumentationSearch term hygiene: a weekly thirty-minute routineThe standing weekly routine we run on every search account: which reports to open, what to negate, what to promote, and what to deliberately leave alone.
- FrameworkAllocating a B2B paid budget across three channelsSearch captures existing demand, professional social creates it, and retargeting only finishes what the other two started. A budget that ignores those roles will always look inefficient somewhere.
- GuideAdvertising on community forums without being rejectedCommunity platforms punish marketing language faster than any other channel. The creative that works reads like a useful comment and the targeting is subject-first, not demographic.
- PlaybookBuilding an account-based audience worth advertising toUploading a list is not targeting. This is how we build, size and refresh account lists so professional social spend reaches the buying committee rather than the whole industry.