Should one ad campaign do all the testing and all the scaling

What this step is

Running two campaigns, testing and scaling, is step 48 (Part VIII, Creative and launch), marked moderate difficulty, often skipped. Running one campaign feels like less admin, right up until that campaign has to serve two incompatible jobs at once.

Why one campaign can't do both jobs

A testing campaign exists to find creatives that convert, and gets judged on hit rate. A scaling campaign exists to spend behind what already works, and gets judged on cost per purchase. A campaign asked to do both does the second badly while looking busy at the first, because new, unproven creative keeps diluting the spend that should be going to a known winner.

Reading delivery in the right order

Read gates in a fixed order and stop at the first one that fails: delivery (impressions, failing under roughly 1,000 because the creative hasn't been served enough to judge), intent (cost per landing page view, where a bad reading implicates the creative), then purchase (cost per purchase against target, where a bad reading with intent fine implicates the page or the offer instead).

Click rate only explains movement inside the first two gates. It says nothing about the third, which is the one that decides whether the spend is working.

A secondary failure mode worth naming

Two ad sets aimed at the same audience will bid against each other. It shows up as rising cost and falling delivery on both sets simultaneously, while total results stay flat, and it's easy to mistake for a market-wide cost increase rather than self-competition.

Where it breaks

Judging a testing campaign on cost per purchase, which kills creatives before they've had enough spend to be judged fairly. Running one campaign for everything because it looks simpler, then reading the resulting flat performance as a market problem rather than a structural one.