Diagnostic signals, evaluation questions, and the operating stack behind the accounts Akorn runs. Field notes from active client work, written by the people doing it.
Five checkable signs your ad reporting is misleading you, from blended conversion counts to systems that disagree while nobody owns the difference.
Six questions that decide whether AI helps your ads operation or just automates its ambiguity: APIs, read vs write, the rulebook, verification, and what stays human.
Five free questions that sort claims from systems: access, audit logs, the rulebook, verification gates, and one live definition test.
Seven production catches invisible on dashboards: blended conversion rollups, renamed campaigns zeroing reports, ghost catalogs, and done that was not done.
metrics.conversions rollups, all_conversions confusion, rename-broken filters, tROAS levels, conversion type mislabels, change_event limits, ghost counts.
Inventory your APIs, connect read-only, write the rulebook, set the cadence, keep judgment human. The step-by-step version, with time estimates.
Feed Google Smart Bidding contribution margin instead of gross revenue so it optimizes to profit. Three steps, plus what the August 17 bidding change means.
Three tools, three different questions. When live queryability beats a pipeline, when it cannot, and the cheap mistake of buying the serious-sounding one.