What evidence should lead a South Africa network choice?
Compare formats, targeting, source transparency, tracking, billing controls and accepted acquisition cost in a matched campaign. Neither reported reach nor a generic best-network label answers that question.
How should the campaign's decision rule be documented?
State the accepted event, maximum cost, eligible settings, destination, attribution window, reporting fields, minimum maturity, loss limit and owner before launch. Reopen the brief after material changes.
What should survive the path from an interaction to reporting?
Campaign and source identifiers should pass through the placement, creative, redirects and destination into the accepted outcome. Review each layer separately when reported totals fail to reconcile.
How can an advertiser test only evidence-ready inventory?
Confirm actual formats, placement contexts, identifiers, caps and exclusions before launch. Design the initial source set around what can be measured and controlled, and record every missing field.
Which creative records help explain a later result?
Keep the concept ID, original asset, format, destination version, message angle and launch date. Those records separate real idea changes from cropping, compression, fatigue or source-mix shifts.
What makes a destination compatible with the campaign promise?
It identifies the advertiser, explains the genuine value and conditions, works on targeted devices and presents a clear action. Track its response and accepted outcomes by source and creative.
Which rules belong in the attribution setup before delivery?
Define the conversion window, duplicate handling, time zone, currency treatment and accepted outcome. Keep test events separate, pass stable identifiers and reconcile platform and backend records on schedule.
Why should source quality be evaluated with several signals?
Behavior, engagement, acceptance, rejection and later value can point to different causes. Investigate tracking and page health first, then document any exclusion and the evidence needed for retesting.
When should exposure be reduced rather than immediately stopped?
Reduce a segment when mature cost or quality is weakening but the cause may be repairable. Pause it when a written loss, tracking, compliance or destination threshold is crossed.
How does a control cohort protect a South Africa scale test?
It preserves the trusted baseline while one bid, budget, source, device, region or creative setting changes. Compare marginal accepted value and restore the earlier setting if quality deteriorates.