ads

Use the Meta decision system playbook, which anchors kill decisions to a target cost per lead with defined spend thresholds and fatigue bands. The Skill routes operational kill/keep/scale questions to that reference rather than giving generic advice.

The Google Search playbook covers an intent ladder, match-type gates, negative keyword strategy, and bidding by conversion volume. It also includes Performance Max guardrails and offline conversion setup for lead-quality feedback.

In the Andromeda era, broad targeting with specific creative typically outperforms interest stacking. The recommended test is duplicating your best ad, removing all interest filters, and comparing CPAs side by side for seven days.

During the first two to four weeks, allocate roughly 70% to proven campaigns and 30% to testing new audiences and creative. When scaling, increase budgets about 20% at a time and wait three to five days between increases.

Yes, but RSA generation must follow the mandatory rsa-output-spec reference, which enforces character limits, required sidecar artifacts like sitelinks and negatives, output ordering, and a pre-send self-check.

It provides strategy, playbooks, and draft assets but does not publish live campaigns; publishing requires human approval in the MyCity workflow. It also prohibits invented metrics or testimonials, requiring real data or explicit placeholders.